2022-10-27 11:11:29 +00:00
// Copyright (c) Facebook, Inc. and its affiliates.
//
// This source code is licensed under the MIT license found in the
2021-10-06 02:16:20 +00:00
// LICENSE file in the root directory of this source tree.
#include <kernels.cuh>
#include <cub/block/block_radix_sort.cuh>
#include <cub/warp/warp_reduce.cuh>
#include <cub/block/block_load.cuh>
#include <cub/block/block_discontinuity.cuh>
#include <cub/block/block_store.cuh>
#include <cub/block/block_reduce.cuh>
#include <cub/cub.cuh>
#include <math_constants.h>
2023-04-26 00:15:51 +00:00
#include <thrust/host_vector.h>
#include <thrust/device_vector.h>
2023-05-01 00:43:02 +00:00
#include <mma.h>
2023-04-26 00:15:51 +00:00
2023-04-27 22:12:49 +00:00
2021-10-06 02:16:20 +00:00
#define HLF_MAX 65504
#define TH 1024
#define NUM 4
#define NUM_BLOCK 4096
2023-05-01 00:43:02 +00:00
2021-10-06 02:16:20 +00:00
// source: https://stackoverflow.com/questions/17399119/how-do-i-use-atomicmax-on-floating-point-values-in-cuda
__device__ float atomicMax(float* address, float val) {
int* address_as_i = reinterpret_cast<int*>(address);
int old = *address_as_i, assumed;
do {
assumed = old;
old = atomicCAS(
reinterpret_cast<int*>(address), assumed,
__float_as_int(fmaxf(val, __int_as_float(assumed))));
} while (assumed != old);
return __int_as_float(old);
}
__device__ float atomicMin(float* address, float val) {
int* address_as_i = reinterpret_cast<int*>(address);
int old = *address_as_i, assumed;
do {
assumed = old;
old = atomicCAS(
reinterpret_cast<int*>(address), assumed,
__float_as_int(fminf(val, __int_as_float(assumed))));
} while (assumed != old);
return __int_as_float(old);
}
2023-02-04 22:52:04 +00:00
__device__ float dDequantizeFP4(unsigned char val, float absmax)
{
float sign = (val & 0b1000) == 8 ? -1.0f : 1.0f;
if((val & 0b0110) == 0)
{
// subnormal
if((val & 0b0001) == 0)
return 0.0f;
else
return sign*0.0625f*absmax;
}
else
{
// normal
float exponent = ((val & 0b0100) == 4 ? 2.0f : 8.0f) + ((val & 0b0010) == 2 ? 0.0f : 2.0f);
float fraction = (val & 0b0001) == 1 ? 1.5f : 1.0f;
return sign*exponent*fraction*absmax;
}
}
2023-04-30 04:52:47 +00:00
__device__ float d2DequantizeFP4(unsigned char val)
{
float sign = (val & 0b1000) == 8 ? -1.0f : 1.0f;
if((val & 0b0110) == 0)
{
// subnormal
if((val & 0b0001) == 0)
return 0.0f;
else
return sign*0.0625f;
}
else
{
// normal
float exponent = ((val & 0b0100) == 4 ? 2.0f : 8.0f) + ((val & 0b0010) == 2 ? 0.0f : 2.0f);
float fraction = (val & 0b0001) == 1 ? 1.5f : 1.0f;
return sign*exponent*fraction;
}
}
2023-04-02 19:42:01 +00:00
__device__ float dDequantizeFP4Tree(unsigned char val, float absmax)
{
float sign = (val & 0b1000) == 8 ? -1.0f : 1.0f;
if((val & 0b0100) == 4) // 0
if((val & 0b0010) == 2) //01
if((val & 0b0001) == 1) // 111
return 0.25000000f*absmax*sign; // 1111
else
return 0.16666667f*absmax*sign; // 1110
else
if((val & 0b0001) == 1) // 110
return 0.50000000f*absmax*sign; // 1101
else
return 0.33333333f*absmax*sign; // 1100
else
if((val & 0b0010) == 2) //10
if((val & 0b0001) == 1) // 101
return 1.00000000f*absmax*sign; // 1011
else
return 0.66666667f*absmax*sign; // 1010
else
if((val & 0b0001) == 1) // 100
return 5.208333333e-03f*absmax*sign; // 1001
else
return 0.00000000f*absmax*sign; // 1000
}
2023-02-04 22:52:04 +00:00
__device__ unsigned char dQuantizeFP4(float x)
{
// FP4 with bias of 3
// first bit is a sign
// subnormals
// 0b000 = 0
// 0b001 = 0.0625
// 0b110 = 2
// 0b111 = 3
// 0b100 = 4
// 0b101 = 6
// 0b010 = 8
// 0b011 = 12
2023-04-02 19:42:01 +00:00
// we do a binary search
// the pivots are divided by 12 (the FP4 absmax)
// since we assum input data is in [-1.0, 1.0]
// !be careful here, its easy to make a mistake
// that is difficult to noice if you add an extra
// zero somewhere!
int sign = x < 0 ? 0b1000 : 0b0000;
x = fabsf(x);
if(x > 0.29166667f)
if( x > 0.583333f)
if( x > 0.8333333f)
return 0b0011+sign;
else
return 0b0010+sign;
else
if(x > 0.4166667f)
return 0b101+sign;
else
return 0b100+sign;
else
if(x > 0.0859375f)
if(x > 0.20833333f)
return 0b0111+sign;
else
return 0b0110+sign;
else
if(x > 0.00260417f)
return 0b0001+sign;
else
return 0b0000+sign;
}
2023-04-30 04:52:47 +00:00
__device__ half dhDequantizeNF4(unsigned char val)
{
// the values for this tree was generated by test_normal_map_tree
// in the file tests/test_functional.py
if((val & 0b1000) == 8)
if((val & 0b0100) == 4) // 1
if((val & 0b0010) == 2) // 11
if((val & 0b0001) == 1) // 111
return 1.0f;
else
return 0.7229568362236023f;
else
if((val & 0b0001) == 1) // 110
return 0.5626170039176941f;
else
return 0.44070982933044434f;
else
if((val & 0b0010) == 2) //10
if((val & 0b0001) == 1) // 101
return 0.33791524171829224f;
else
return 0.24611230194568634f;
else
if((val & 0b0001) == 1) // 100
return 0.16093020141124725f;
else
return 0.07958029955625534f;
else
if((val & 0b0100) == 4) // 0
if((val & 0b0010) == 2) //01
if((val & 0b0001) == 1) // 011
return 0.0f;
else
return -0.09105003625154495f;
else
if((val & 0b0001) == 1) // 010
return -0.18477343022823334f;
else
return -0.28444138169288635f;
else
if((val & 0b0010) == 2) //00
if((val & 0b0001) == 1) // 001
return -0.39491748809814453f;
else
return -0.5250730514526367f;
else
if((val & 0b0001) == 1) // 000
return -0.6961928009986877f;
else
return -1.0f;
}
__device__ float dDequantizeNF4(unsigned char val)
2023-04-02 23:10:35 +00:00
{
2023-07-11 12:55:49 +00:00
2023-04-02 23:10:35 +00:00
// the values for this tree was generated by test_normal_map_tree
// in the file tests/test_functional.py
if((val & 0b1000) == 8)
if((val & 0b0100) == 4) // 1
if((val & 0b0010) == 2) // 11
if((val & 0b0001) == 1) // 111
2023-04-30 04:52:47 +00:00
return 1.0f;
2023-04-02 23:10:35 +00:00
else
2023-04-30 04:52:47 +00:00
return 0.7229568362236023f;
2023-04-02 23:10:35 +00:00
else
if((val & 0b0001) == 1) // 110
2023-04-30 04:52:47 +00:00
return 0.5626170039176941f;
2023-04-02 23:10:35 +00:00
else
2023-04-30 04:52:47 +00:00
return 0.44070982933044434f;
2023-04-02 23:10:35 +00:00
else
if((val & 0b0010) == 2) //10
if((val & 0b0001) == 1) // 101
2023-04-30 04:52:47 +00:00
return 0.33791524171829224f;
2023-04-02 23:10:35 +00:00
else
2023-04-30 04:52:47 +00:00
return 0.24611230194568634f;
2023-04-02 23:10:35 +00:00
else
if((val & 0b0001) == 1) // 100
2023-04-30 04:52:47 +00:00
return 0.16093020141124725f;
2023-04-02 23:10:35 +00:00
else
2023-04-30 04:52:47 +00:00
return 0.07958029955625534f;
2023-04-02 23:10:35 +00:00
else
if((val & 0b0100) == 4) // 0
if((val & 0b0010) == 2) //01
if((val & 0b0001) == 1) // 011
2023-04-30 04:52:47 +00:00
return 0.0f;
2023-04-02 23:10:35 +00:00
else
2023-04-30 04:52:47 +00:00
return -0.09105003625154495f;
2023-04-02 23:10:35 +00:00
else
if((val & 0b0001) == 1) // 010
2023-04-30 04:52:47 +00:00
return -0.18477343022823334f;
2023-04-02 23:10:35 +00:00
else
2023-04-30 04:52:47 +00:00
return -0.28444138169288635f;
2023-04-02 23:10:35 +00:00
else
if((val & 0b0010) == 2) //00
if((val & 0b0001) == 1) // 001
2023-04-30 04:52:47 +00:00
return -0.39491748809814453f;
2023-04-02 23:10:35 +00:00
else
2023-04-30 04:52:47 +00:00
return -0.5250730514526367f;
2023-04-02 23:10:35 +00:00
else
if((val & 0b0001) == 1) // 000
2023-04-30 04:52:47 +00:00
return -0.6961928009986877f;
2023-04-02 23:10:35 +00:00
else
2023-04-30 04:52:47 +00:00
return -1.0f;
2023-04-02 23:10:35 +00:00
}
2023-04-03 18:00:12 +00:00
__device__ unsigned char dQuantizeNF4(float x)
2023-04-02 19:42:01 +00:00
{
2023-04-02 23:10:35 +00:00
// the values for this tree was generated by test_normal_map_tree
// in the file tests/test_functional.py
if(x > 0.03979014977812767f)
if(x > 0.3893125355243683f) // 1
if(x > 0.6427869200706482f) // 11
if(x > 0.8614784181118011f) // 111
return 0b1111;
else
return 0b1110;
2023-02-04 22:52:04 +00:00
else
2023-04-02 23:10:35 +00:00
if(x > 0.5016634166240692f) // 110
return 0b1101;
else
return 0b1100;
2023-02-04 22:52:04 +00:00
else
2023-04-02 23:10:35 +00:00
if(x > 0.2035212516784668f) // 10
if(x > 0.2920137718319893f) // 101
return 0b1011;
else
return 0b1010;
2023-02-04 22:52:04 +00:00
else
2023-04-02 23:10:35 +00:00
if(x > 0.1202552504837513f) // 100
return 0b1001;
else
2023-04-03 18:00:12 +00:00
return 0b1000;
2023-02-04 22:52:04 +00:00
else
2023-04-02 23:10:35 +00:00
if(x > -0.33967943489551544f) // 0
if(x > -0.13791173323988914f) // 01
if(x > -0.045525018125772476f) // 011
return 0b0111;
else
return 0b0110;
2023-02-04 22:52:04 +00:00
else
2023-04-02 23:10:35 +00:00
if(x > -0.23460740596055984f) // 010
return 0b0101;
else
return 0b0100;
2023-02-04 22:52:04 +00:00
else
2023-04-02 23:10:35 +00:00
if(x > -0.6106329262256622f) // 00
if(x > -0.4599952697753906f) // 001
return 0b0011;
else
return 0b0010;
2023-02-04 22:52:04 +00:00
else
2023-04-02 23:10:35 +00:00
if(x > -0.8480964004993439f) // 000
return 0b0001;
else
return 0b0000;
2023-02-04 22:52:04 +00:00
}
2023-03-09 19:54:54 +00:00
// sign function for lion
// taken from https://stackoverflow.com/a/4609795, but not sure if there's a proper way to do this in CUDA
2023-05-07 20:34:03 +00:00
template <typename T> __device__ int sgn(T val)
{
2023-03-09 19:54:54 +00:00
return (T(0) < val) - (val < T(0));
}
2023-02-04 22:52:04 +00:00
2021-10-06 02:16:20 +00:00
template <int STOCHASTIC>
__device__ unsigned char dQuantize(float* smem_code, const float rand, float x)
{
int pivot = 127;
int upper_pivot = 255;
int lower_pivot = 0;
float lower = -1.0f;
float upper = 1.0f;
float val = smem_code[pivot];
// i>>=1 = {32, 16, 8, 4, 2, 1}
for(int i = 64; i > 0; i>>=1)
{
if(x > val)
{
lower_pivot = pivot;
lower = val;
pivot+=i;
}
else
{
upper_pivot = pivot;
upper = val;
pivot-=i;
}
val = smem_code[pivot];
}
if(upper_pivot == 255)
upper = smem_code[upper_pivot];
if(lower_pivot == 0)
lower = smem_code[lower_pivot];
if(!STOCHASTIC)
{
if(x > val)
{
float midpoint = (upper+val)*0.5f;
if(x > midpoint)
{
return upper_pivot;
}
else
return pivot;
}
else
{
float midpoint = (lower+val)*0.5f;
if(x < midpoint)
return lower_pivot;
else
return pivot;
}
}
else
{
if(x > val)
{
float dist_to_upper = fabsf(upper-x);
float dist_full = upper-val;
if(rand >= dist_to_upper/dist_full) return upper_pivot;
else return pivot;
}
else
{
float dist_to_lower = fabsf(lower-x);
float dist_full = val-lower;
if(rand >= dist_to_lower/dist_full) return lower_pivot;
else return pivot;
}
}
}
template <int SIGNED>
__device__ __forceinline__ unsigned char quantize_2D(float *__restrict__ quadrants, float *__restrict__ const smem_code, float x)
{
int pivot = 127;
int upper_pivot = 255;
int lower_pivot = 0;
float lower = SIGNED ? -1.0f : 0.0f;
float upper = 1.0f;
float midpoint;
float val = quadrants[1];
int local_pivot = 1;
int offset = 1;
// i>>=1 = {32, 16, 8, 4, 2, 1}
for(int i = 64; i > 0; i>>=1)
{
if(x > val)
{
lower_pivot = pivot;
lower = val;
pivot+=i;
//val = i == 64 ? quadrants[2] : smem_code[pivot];
local_pivot += offset;
}
else
{
upper_pivot = pivot;
upper = val;
pivot-=i;
//val = i == 64 ? quadrants[0] : smem_code[pivot];
local_pivot -= offset;
}
val = i >= 64 ? quadrants[local_pivot] : smem_code[pivot];
offset -= 1;
}
if(x > val)
{
midpoint = (upper+val)*0.5f;
if(x > midpoint)
return upper_pivot;
else
return pivot;
}
else
{
midpoint = (lower+val)*0.5f;
if(x < midpoint)
return lower_pivot;
else
return pivot;
}
}
template <int SIGNED>
__device__ __forceinline__ unsigned char quantize_quadrant(int QUADRANT, float *__restrict__ const smem_code, float x, float lower, float midpoint, float upper)
{
int lower_pivot = QUADRANT*16-1 - 0;
int pivot = QUADRANT*16-1 + 16;
int upper_pivot = QUADRANT*16-1 + 31;
float val = midpoint;
// i>>=1 = {32, 16, 8, 4, 2, 1}
for(int i = 16; i > 0; i>>=1)
{
if(x > val)
{
lower_pivot = pivot;
lower = val;
pivot+=i;
}
else
{
upper_pivot = pivot;
upper = val;
pivot-=i;
}
val = smem_code[pivot];
}
if(x > val)
{
midpoint = (upper+val)*0.5f;
if(x > midpoint)
return upper_pivot;
else
return pivot;
}
else
{
midpoint = (lower+val)*0.5f;
if(x < midpoint)
return lower_pivot;
else
return pivot;
}
}
__global__ void kHistogramScatterAdd2D(float* histogram, int *index1, int *index2, float *src, const int maxidx1, const int n)
{
const int tid = threadIdx.x + (blockDim.x*blockIdx.x);
const int numThreads = blockDim.x*gridDim.x;
for(int i = tid; i < n; i+=numThreads)
{
int idx = (index1[i]*maxidx1) + index2[i];
atomicAdd(&histogram[idx], src[i]);
}
}
template<typename T, int BLOCK_SIZE, int NUM_MAX>
__global__ void kCompressMax(T * __restrict__ const A, T* out, unsigned char* out_idx, const int n)
{
typedef cub::WarpReduce<T> WarpReduce;
__shared__ typename WarpReduce::TempStorage temp_storage;
typedef cub::BlockLoad<T, BLOCK_SIZE/8 , 8, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
__shared__ typename LoadT::TempStorage loadt;
const int warp_idx = threadIdx.x/32;
const int valid_items = n - (blockIdx.x*BLOCK_SIZE) > BLOCK_SIZE ? BLOCK_SIZE : n - (blockIdx.x*BLOCK_SIZE);
// BLOCK_SIZE/32 == number of warps
__shared__ int smem_max_indices[8*BLOCK_SIZE/32];
__shared__ float smem_max_values[8*BLOCK_SIZE/32];
T values[8];
T max1 = -64000.0f;
T max2 = -64000.0f;
int max_idx1 = -1;
int max_idx2 = -1;
int sign1 = -1;
int sign2 = -1;
// 1. load 8 values per thread
// 2. compute 2-max in registers (64 max per warp)
// 3. do warp reduction + broadcast back
// 4. Up-shift maxed value, write index into shared memory, replace with 2nd largest
// 5. Repeat (3) 8 times for top 8 values in 256
// 6. store with byte index
LoadT(loadt).Load(&(A[(blockIdx.x*BLOCK_SIZE)]), values, valid_items, (T)0.0f);
#pragma unroll 8
for(int i = 0; i < 8; i++)
{
T absval = fabsf(values[i]);
if(absval > max1)
{
max1 = values[i];
sign1 = signbit(values[i]);
max_idx1 = 8*threadIdx.x + i;
}
else if(absval > max2)
{
max2 = values[i];
sign2 = signbit(values[i]);
max_idx2 = 8*threadIdx.x + i;
}
}
float warp_max;
for(int i = 0; i < 8; i++)
{
// 3. do warp reduction + broadcast back
warp_max = WarpReduce(temp_storage).Reduce(max1, cub::Max());
warp_max = cub::ShuffleIndex<32>(warp_max, 0, 0xffffffff);
// 4. Up-shift maxed value, write index into shared memory, replace with 2nd largest
if(warp_max == max1)
{
smem_max_values[warp_idx*8 + i] = sign1 != 0 ? -max1 : max1;
smem_max_indices[warp_idx*8 + i] = max_idx1;
sign1 = sign2;
max1 = max2;
max_idx1 = max_idx2;
max2 = -64000.0f;
}
__syncwarp();
}
if(threadIdx.x % 32 < 8)
{
// offset: 8 values per 256 input values
2022-10-27 11:11:29 +00:00
//
2021-10-06 02:16:20 +00:00
int offset = BLOCK_SIZE*blockIdx.x*BLOCK_SIZE/32*8;
}
}
#define THREADS_ESTIMATE 512
#define NUM_ESTIMATE 8
#define BLOCK_ESTIMATE 4096
template<typename T>
__launch_bounds__(THREADS_ESTIMATE, 1)
__global__ void kEstimateQuantiles(T *__restrict__ const A, float *code, const float offset, const T max_val, const int n)
{
const int n_full = (BLOCK_ESTIMATE*(n/BLOCK_ESTIMATE)) + (n % BLOCK_ESTIMATE == 0 ? 0 : BLOCK_ESTIMATE);
int valid_items = (blockIdx.x+1 == gridDim.x) ? n - (blockIdx.x*BLOCK_ESTIMATE) : BLOCK_ESTIMATE;
const int base_idx = (blockIdx.x * BLOCK_ESTIMATE);
const float reciprocal_num_blocks = 1.0f/(n < 4096 ? 1.0f : (n/BLOCK_ESTIMATE));
T vals[NUM_ESTIMATE];
typedef cub::BlockRadixSort<T, THREADS_ESTIMATE, NUM_ESTIMATE, cub::NullType, 4, true, cub::BLOCK_SCAN_RAKING> BlockRadixSort;
typedef cub::BlockLoad<T, THREADS_ESTIMATE, NUM_ESTIMATE, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadFloat;
__shared__ union {
typename LoadFloat::TempStorage loadf;
typename BlockRadixSort::TempStorage sort;
int smem_qidx[BLOCK_ESTIMATE];
} temp_storage;
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*BLOCK_ESTIMATE)
{
valid_items = n - i > BLOCK_ESTIMATE ? BLOCK_ESTIMATE : n - i;
// do not process half-blocks
if(valid_items < BLOCK_ESTIMATE && n > BLOCK_ESTIMATE){ continue; }
#pragma unroll 4
for(int j = 0; j < NUM_ESTIMATE; j++)
vals[j] = max_val;
__syncthreads();
LoadFloat(temp_storage.loadf).Load(&(A[i]), vals, valid_items);
#pragma unroll 4
for(int j = 0; j < NUM_ESTIMATE; j++)
vals[j] = ((float)vals[j]) * reciprocal_num_blocks;
__syncthreads();
// sort into striped pattern to mitigate bank conflicts
// striped pattern index for thread 0 [0, 1024, 2048, 3096]
// striped pattern index for thread 1 [1, 1025, 2049, 3097]
BlockRadixSort(temp_storage.sort).SortBlockedToStriped(vals);
__syncthreads();
for(int j = threadIdx.x; j < BLOCK_ESTIMATE; j+=blockDim.x)
temp_storage.smem_qidx[j] = -1;
if(threadIdx.x < 256)
{
float q_interval = (1.0f-(2.0f*offset))/255.0f;
int local_idx = round(((offset+(threadIdx.x*q_interval))*(valid_items-1)));
temp_storage.smem_qidx[local_idx] = threadIdx.x;
}
__syncthreads();
for(int i = threadIdx.x; i < BLOCK_ESTIMATE; i+=blockDim.x)
{
if(temp_storage.smem_qidx[i] != -1)
atomicAdd(&code[temp_storage.smem_qidx[i]], vals[i/THREADS_ESTIMATE]);
}
}
}
__launch_bounds__(TH, 4)
__global__ void kQuantize(float * code, float * __restrict__ const A, unsigned char *out, const int n)
{
const int n_full = (NUM_BLOCK*(n/NUM_BLOCK)) + (n % NUM_BLOCK == 0 ? 0 : NUM_BLOCK);
int valid_items = (blockIdx.x+1 == gridDim.x) ? n - (blockIdx.x*NUM_BLOCK) : NUM_BLOCK;
const int base_idx = (blockIdx.x * NUM_BLOCK);
float vals[NUM];
unsigned char qvals[NUM];
//const int lane_id = threadIdx.x % 2;
typedef cub::BlockLoad<float, TH, NUM, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadFloat;
typedef cub::BlockStore<unsigned char, TH, NUM, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreChar;
__shared__ typename LoadFloat::TempStorage loadf;
__shared__ typename StoreChar::TempStorage storec;
__shared__ float smem_code[256];
//__shared__ float smem_code[2][257];
if(threadIdx.x < 256)
{
smem_code[threadIdx.x] = code[threadIdx.x];
//smem_code[0][threadIdx.x] = code[threadIdx.x];
//smem_code[1][threadIdx.x] = smem_code[0][threadIdx.x];
}
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*NUM_BLOCK)
{
// number of values already processed in blocks +
// number of values already processed in this block +
// rand_offset % mod value
valid_items = n - i > NUM_BLOCK ? NUM_BLOCK : n - i;
__syncthreads();
LoadFloat(loadf).Load(&(A[i]), vals, valid_items);
#pragma unroll 4
for(int j = 0; j < NUM; j++)
qvals[j] = dQuantize<0>(smem_code, 0.0f, vals[j]);
__syncthreads();
StoreChar(storec).Store(&(out[i]), qvals, valid_items);
}
}
2023-04-02 23:10:35 +00:00
template<typename T, int BLOCK_SIZE, int NUM_PER_TH, int STOCHASTIC, int DATA_TYPE>
2022-11-07 00:27:48 +00:00
//__launch_bounds__(TH, 4)
2021-10-06 02:16:20 +00:00
__global__ void kQuantizeBlockwise(float * code, T * __restrict__ const A, float *absmax, unsigned char *out, float * __restrict__ const rand, const int rand_offset, const int n)
{
const int n_full = gridDim.x * BLOCK_SIZE;
int valid_items = 0;
const int base_idx = (blockIdx.x * BLOCK_SIZE);
2022-11-07 00:27:48 +00:00
T vals[NUM_PER_TH];
float rand_vals[NUM_PER_TH];
2023-04-02 23:10:35 +00:00
unsigned char qvals[(DATA_TYPE > 0) ? NUM_PER_TH/2 : NUM_PER_TH];
2021-10-06 02:16:20 +00:00
//float local_abs_max = -FLT_MAX;
float local_abs_max = 0.0f;
int local_rand_idx = 0;
typedef cub::BlockLoad<T, BLOCK_SIZE/NUM_PER_TH, NUM_PER_TH, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
2023-04-02 23:10:35 +00:00
typedef cub::BlockStore<unsigned char, BLOCK_SIZE/NUM_PER_TH, (DATA_TYPE > 0) ? NUM_PER_TH/2 : NUM_PER_TH, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreChar;
2021-10-06 02:16:20 +00:00
typedef cub::BlockReduce<float, BLOCK_SIZE/NUM_PER_TH> BlockReduce;
typedef cub::BlockLoad<float, BLOCK_SIZE/NUM_PER_TH, NUM_PER_TH, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadFloat;
__shared__ typename LoadT::TempStorage loadt;
__shared__ typename LoadFloat::TempStorage loadf;
__shared__ typename StoreChar::TempStorage storec;
__shared__ typename BlockReduce::TempStorage reduce;
__shared__ float smem_code[256];
__shared__ float smem_absmax_value[1];
2023-04-02 23:10:35 +00:00
if(DATA_TYPE == General8bit)
2023-02-04 22:52:04 +00:00
for(int i = threadIdx.x; i < 256; i+=blockDim.x)
smem_code[i] = code[i];
2021-10-06 02:16:20 +00:00
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*BLOCK_SIZE)
{
valid_items = n - i > BLOCK_SIZE ? BLOCK_SIZE : n - i;
local_abs_max = -FLT_MAX;
__syncthreads();
LoadT(loadt).Load(&(A[i]), vals, valid_items, (T)0.0f);
// 1. compute local max
// 2. broadcast local max
// 3. normalize inputs and quantize
#pragma unroll NUM_PER_TH
for(int j = 0; j < NUM_PER_TH; j++)
local_abs_max = fmaxf(local_abs_max, fabsf((float)vals[j]));
local_abs_max = BlockReduce(reduce).Reduce(local_abs_max, cub::Max(), valid_items);
if(threadIdx.x == 0)
smem_absmax_value[0] = local_abs_max;
__syncthreads();
if(threadIdx.x == 0)
absmax[i/BLOCK_SIZE] = local_abs_max;
else
local_abs_max = smem_absmax_value[0];
__syncwarp();
local_abs_max = 1.0f/local_abs_max;
if(STOCHASTIC)
{
local_rand_idx = ((blockIdx.x*NUM_BLOCK) + (threadIdx.x*NUM) + rand_offset) % (1024-4);
LoadFloat(loadf).Load(&rand[local_rand_idx], rand_vals, BLOCK_SIZE, 0);
}
2023-04-02 23:10:35 +00:00
unsigned char packed_4bit = 0;
switch(DATA_TYPE)
2021-10-06 02:16:20 +00:00
{
2023-04-02 23:10:35 +00:00
case General8bit:
#pragma unroll NUM_PER_TH
for(int j = 0; j < NUM_PER_TH; j++)
{
if(!STOCHASTIC)
qvals[j] = dQuantize<0>(smem_code, 0.0f, ((float)vals[j])*local_abs_max);
else
qvals[j] = dQuantize<1>(smem_code, rand_vals[j], ((float)vals[j])*local_abs_max);
}
break;
case FP4:
#pragma unroll NUM_PER_TH
for(int j = 0; j < NUM_PER_TH/2; j++)
{
packed_4bit |= dQuantizeFP4(((float)vals[2*j])*local_abs_max) << 4;
packed_4bit |= dQuantizeFP4(((float)vals[2*j+1])*local_abs_max);
qvals[j] = packed_4bit;
}
break;
case NF4:
#pragma unroll NUM_PER_TH
for(int j = 0; j < NUM_PER_TH/2; j++)
{
2023-04-03 18:00:12 +00:00
packed_4bit |= dQuantizeNF4(((float)vals[2*j])*local_abs_max) << 4;
packed_4bit |= dQuantizeNF4(((float)vals[2*j+1])*local_abs_max);
2023-04-02 23:10:35 +00:00
qvals[j] = packed_4bit;
}
break;
2021-10-06 02:16:20 +00:00
}
__syncthreads();
2023-04-02 23:10:35 +00:00
StoreChar(storec).Store(&(out[(DATA_TYPE > 0) ? i/2 : i]), qvals, (DATA_TYPE > 0) ? (valid_items+1)/2 : valid_items);
2021-10-06 02:16:20 +00:00
}
}
2023-04-03 18:00:12 +00:00
template<typename T, int TILE_SIZE, int THREADS, int NUM_PER_TH, int DATA_TYPE>
2023-02-04 22:52:04 +00:00
__global__ void kDequantizeBlockwise(float *code, unsigned char * A, float * absmax, T *out, const int blocksize, const int n)
2021-10-06 02:16:20 +00:00
{
2023-02-04 22:52:04 +00:00
const int n_load = (gridDim.x * TILE_SIZE);
int valid_items_load = 0;
int valid_items_store = 0;
const int base_idx = (blockIdx.x * TILE_SIZE);
2021-10-06 02:16:20 +00:00
2023-04-03 18:00:12 +00:00
T vals[NUM_PER_TH*((DATA_TYPE > 0) ? 2 : 1)];
2022-11-07 00:27:48 +00:00
unsigned char qvals[NUM_PER_TH];
2021-10-06 02:16:20 +00:00
float local_abs_max = -FLT_MAX;
typedef cub::BlockLoad<unsigned char, THREADS, NUM_PER_TH, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadChar;
2023-04-03 18:00:12 +00:00
typedef cub::BlockStore<T, THREADS, NUM_PER_TH*((DATA_TYPE > 0) ? 2 : 1), cub::BLOCK_STORE_WARP_TRANSPOSE> StoreT;
2021-10-06 02:16:20 +00:00
__shared__ typename LoadChar::TempStorage loadchar;
__shared__ typename StoreT::TempStorage storet;
2023-02-04 22:52:04 +00:00
for (unsigned int i = base_idx; i < n_load; i += gridDim.x*TILE_SIZE)
2021-10-06 02:16:20 +00:00
{
2023-04-03 18:00:12 +00:00
if(DATA_TYPE > 0)
{
valid_items_load = (n+1)/2 - i > TILE_SIZE ? TILE_SIZE : (n+1)/2 - i;
valid_items_store = n - i*2 > TILE_SIZE*2 ? TILE_SIZE*2 : n - i*2;
}
else
{
valid_items_load = n - i > TILE_SIZE ? TILE_SIZE : n - i;
valid_items_store = n - i > TILE_SIZE ? TILE_SIZE : n - i;
}
local_abs_max = __ldg(&absmax[(i+threadIdx.x*NUM_PER_TH)/(blocksize)]);
2021-10-06 02:16:20 +00:00
2023-04-03 18:00:12 +00:00
__syncthreads();
LoadChar(loadchar).Load(&(A[i]), qvals, valid_items_load, 128);
2021-10-06 02:16:20 +00:00
2023-04-03 18:00:12 +00:00
switch(DATA_TYPE)
{
case General8bit:
// load code through read-only cache via __ldg
#pragma unroll NUM_PER_TH
for(int j = 0; j < NUM_PER_TH; j++)
vals[j] = __ldg(&code[qvals[j]])*local_abs_max;
break;
case FP4:
#pragma unroll NUM_PER_TH
for(int j = 0; j < NUM_PER_TH; j++)
{
vals[j*2] = dDequantizeFP4Tree(qvals[j] >> 4, local_abs_max);
vals[j*2 + 1] = dDequantizeFP4Tree(qvals[j] & 0x0F, local_abs_max);
}
break;
case NF4:
#pragma unroll NUM_PER_TH
for(int j = 0; j < NUM_PER_TH; j++)
{
2023-04-30 04:52:47 +00:00
vals[j*2] = dDequantizeNF4(qvals[j] >> 4)* local_abs_max;
vals[j*2 + 1] = dDequantizeNF4(qvals[j] & 0x0F)* local_abs_max;
2023-04-03 18:00:12 +00:00
}
break;
}
2021-10-06 02:16:20 +00:00
2023-04-03 18:00:12 +00:00
__syncthreads();
StoreT(storet).Store(&(out[(DATA_TYPE > 0) ? i*2 : i]), vals, valid_items_store);
2021-10-06 02:16:20 +00:00
}
}
__global__ void kDequantize(float *code, unsigned char *A, float *out, const int n)
{
const unsigned int numThreads = blockDim.x * gridDim.x;
const int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
__shared__ float smem_code[256];
if(threadIdx.x < 256)
{
smem_code[threadIdx.x] = code[threadIdx.x];
}
__syncthreads();
for (int i = idx;i < n; i += numThreads)
{
out[i] = smem_code[A[i]];
}
}
template<typename T, int OPTIMIZER, int BLOCK_SIZE, int NUM_VALS>
__launch_bounds__(BLOCK_SIZE/NUM_VALS, 1)
2022-10-27 11:11:29 +00:00
__global__ void kPreconditionOptimizer32bit2State(T* g, T* p,
2021-10-06 02:16:20 +00:00
float* state1, float* state2, float *unorm,
const float beta1, const float beta2, const float eps, const float weight_decay,
const int step, const float lr, const float gnorm_scale, const int n)
{
const int n_full = (BLOCK_SIZE*(n/BLOCK_SIZE)) + (n % BLOCK_SIZE == 0 ? 0 : BLOCK_SIZE);
const int base_idx = (blockIdx.x * blockDim.x * NUM_VALS);
int valid_items = 0;
T g_vals[NUM_VALS];
float s1_vals[NUM_VALS];
float s2_vals[NUM_VALS];
const float correction1 = 1.0f/(1.0f - powf(beta1, step));
const float correction2 = 1.0f/(1.0f - powf(beta2, step));
typedef cub::BlockLoad<T, BLOCK_SIZE/NUM_VALS, NUM_VALS, cub::BLOCK_LOAD_WARP_TRANSPOSE> Load;
typedef cub::BlockLoad<float, BLOCK_SIZE/NUM_VALS, NUM_VALS, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadFloat;
typedef cub::BlockReduce<float, BLOCK_SIZE/NUM_VALS> BlockReduce;
__shared__ union {
typename Load::TempStorage load;
typename LoadFloat::TempStorage loadf;
typename BlockReduce::TempStorage reduce;
} temp_storage;
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*BLOCK_SIZE)
{
valid_items = n - i >= (BLOCK_SIZE) ? (BLOCK_SIZE) : n - i;
__syncthreads();
Load(temp_storage.load).Load(&(g[i]), g_vals, valid_items, 0.0f);
__syncthreads();
LoadFloat(temp_storage.loadf).Load(&(state1[i]), s1_vals, valid_items, 0.0f);
__syncthreads();
LoadFloat(temp_storage.loadf).Load(&(state2[i]), s2_vals, valid_items, 0.0f);
# pragma unroll NUM_VALS
for(unsigned int j = 0; j < NUM_VALS; j++)
g_vals[j] = gnorm_scale*((float)g_vals[j]);
# pragma unroll NUM_VALS
for(unsigned int j = 0; j < NUM_VALS; j++)
{
switch(OPTIMIZER)
{
2022-10-27 11:11:29 +00:00
case ADAM:
2021-10-06 02:16:20 +00:00
s1_vals[j] = s1_vals[j]*beta1 + ((1.0f -beta1)*((float)g_vals[j]));
s2_vals[j] = s2_vals[j]*beta2 + ((1.0f -beta2)*(((float)g_vals[j])*((float)g_vals[j])));
s1_vals[j] *= correction1;
s2_vals[j] *= correction2;
s1_vals[j] = s1_vals[j]/(sqrtf(s2_vals[j])+eps); // update
s1_vals[j] *= s1_vals[j]; // update l2 norm (update*update)
break;
}
}
# pragma unroll NUM_VALS-1
for(unsigned int j = 1; j < NUM_VALS; j++)
s1_vals[0] += s1_vals[j];
__syncthreads();
s1_vals[0] = BlockReduce(temp_storage.reduce).Sum(s1_vals[0]);
if(threadIdx.x == 0)
atomicAdd(&unorm[0], s1_vals[0]);
__syncwarp();
}
}
#define NUM_PER_THREAD 4
template<typename T, int OPTIMIZER>
__launch_bounds__(TH, 1)
2022-10-27 11:11:29 +00:00
__global__ void kOptimizer32bit2State(T* g, T* p,
2021-10-06 02:16:20 +00:00
float* state1, float* state2, float *unorm, const float max_unorm, const float param_norm,
const float beta1, const float beta2, const float eps, const float weight_decay,
2021-10-21 01:37:44 +00:00
const int step, const float lr, const float gnorm_scale, const bool skip_zeros, const int n)
2021-10-06 02:16:20 +00:00
{
const int n_full = ((TH*NUM_PER_THREAD)*(n/(TH*NUM_PER_THREAD))) + (n % (TH*NUM_PER_THREAD) == 0 ? 0 : (TH*NUM_PER_THREAD));
const int base_idx = (blockIdx.x * blockDim.x * NUM_PER_THREAD);
int valid_items = 0;
float update_scale = 0.0f;
T g_vals[NUM_PER_THREAD];
T p_vals[NUM_PER_THREAD];
float s1_vals[NUM_PER_THREAD];
float s2_vals[NUM_PER_THREAD];
const float correction1 = 1.0f - powf(beta1, step);
const float correction2 = sqrtf(1.0f - powf(beta2, step));
const float step_size = -lr*correction2/correction1;
if(max_unorm > 0.0f)
{
update_scale = max_unorm > 0.0f ? sqrtf(unorm[0]) : 1.0f;
if(update_scale > max_unorm*param_norm){ update_scale = (max_unorm*param_norm)/update_scale; }
else{ update_scale = 1.0f; }
}
else{ update_scale = 1.0f; }
typedef cub::BlockLoad<T, TH, NUM_PER_THREAD, cub::BLOCK_LOAD_WARP_TRANSPOSE> Load;
typedef cub::BlockStore<T, TH, NUM_PER_THREAD, cub::BLOCK_STORE_WARP_TRANSPOSE> Store;
typedef cub::BlockLoad<float, TH, NUM_PER_THREAD, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadFloat;
typedef cub::BlockStore<float, TH, NUM_PER_THREAD, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreFloat;
__shared__ union {
typename Load::TempStorage load;
typename Store::TempStorage store;
typename LoadFloat::TempStorage loadf;
typename StoreFloat::TempStorage storef;
} temp_storage;
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*TH*NUM_PER_THREAD)
{
valid_items = n - i >= (TH*NUM_PER_THREAD) ? (TH*NUM_PER_THREAD) : n - i;
__syncthreads();
Load(temp_storage.load).Load(&(g[i]), g_vals, valid_items);
__syncthreads();
LoadFloat(temp_storage.loadf).Load(&(state1[i]), s1_vals, valid_items);
__syncthreads();
LoadFloat(temp_storage.loadf).Load(&(state2[i]), s2_vals, valid_items);
__syncthreads();
Load(temp_storage.load).Load(&(p[i]), p_vals, valid_items);
# pragma unroll 4
for(unsigned int j = 0; j < NUM_PER_THREAD; j++)
g_vals[j] = gnorm_scale*((float)g_vals[j]);
# pragma unroll 4
for(unsigned int j = 0; j < NUM_PER_THREAD; j++)
{
switch(OPTIMIZER)
{
2022-10-27 11:11:29 +00:00
case ADAM:
2021-10-22 00:22:43 +00:00
if(!skip_zeros || (skip_zeros && ((float)g_vals[j] != 0.0f)))
2021-10-21 02:15:47 +00:00
{
s1_vals[j] = s1_vals[j]*beta1 + ((1.0f -beta1)*((float)g_vals[j]));
s2_vals[j] = s2_vals[j]*beta2 + ((1.0f -beta2)*(((float)g_vals[j])*((float)g_vals[j])));
p_vals[j] = ((float)p_vals[j]) + (update_scale*step_size*(s1_vals[j]/(sqrtf(s2_vals[j])+(eps*correction2))));
2021-11-29 05:18:11 +00:00
if(weight_decay > 0.0f)
p_vals[j] = ((float)p_vals[j])*(1.0f-(lr*weight_decay));
2021-10-21 02:15:47 +00:00
}
2021-10-06 02:16:20 +00:00
break;
}
}
__syncthreads();
Store(temp_storage.store).Store(&(p[i]), p_vals, valid_items);
__syncthreads();
StoreFloat(temp_storage.storef).Store(&(state1[i]), s1_vals, valid_items);
__syncthreads();
StoreFloat(temp_storage.storef).Store(&(state2[i]), s2_vals, valid_items);
}
}
template<typename T, int OPTIMIZER, int BLOCK_SIZE, int NUM_VALS>
__launch_bounds__(BLOCK_SIZE/NUM_VALS, 1)
2022-10-27 11:11:29 +00:00
__global__ void kPreconditionOptimizer32bit1State(T* g, T* p,
2021-10-06 02:16:20 +00:00
float* state1, float *unorm,
2023-03-10 21:00:59 +00:00
const float beta1, const float beta2, const float eps, const float weight_decay,
2021-10-06 02:16:20 +00:00
const int step, const float lr, const float gnorm_scale, const int n)
{
const int n_full = (BLOCK_SIZE*(n/BLOCK_SIZE)) + (n % BLOCK_SIZE == 0 ? 0 : BLOCK_SIZE);
const int base_idx = (blockIdx.x * blockDim.x * NUM_VALS);
int valid_items = 0;
T g_vals[NUM_VALS];
float s1_vals[NUM_VALS];
typedef cub::BlockLoad<T, BLOCK_SIZE/NUM_VALS, NUM_VALS, cub::BLOCK_LOAD_WARP_TRANSPOSE> Load;
typedef cub::BlockLoad<float, BLOCK_SIZE/NUM_VALS, NUM_VALS, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadFloat;
typedef cub::BlockReduce<float, BLOCK_SIZE/NUM_VALS> BlockReduce;
__shared__ union {
typename Load::TempStorage load;
typename LoadFloat::TempStorage loadf;
typename BlockReduce::TempStorage reduce;
} temp_storage;
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*BLOCK_SIZE)
{
valid_items = n - i >= (BLOCK_SIZE) ? (BLOCK_SIZE) : n - i;
__syncthreads();
Load(temp_storage.load).Load(&(g[i]), g_vals, valid_items, 0.0f);
__syncthreads();
LoadFloat(temp_storage.loadf).Load(&(state1[i]), s1_vals, valid_items, 0.0f);
# pragma unroll NUM_VALS
for(unsigned int j = 0; j < NUM_VALS; j++)
g_vals[j] = gnorm_scale*((float)g_vals[j]);
# pragma unroll NUM_VALS
for(unsigned int j = 0; j < NUM_VALS; j++)
{
switch(OPTIMIZER)
{
2022-10-27 11:11:29 +00:00
case MOMENTUM:
2021-10-06 02:16:20 +00:00
if(step == 1)
s1_vals[j] = (float)g_vals[j]; // state update
else
s1_vals[j] = s1_vals[j]*beta1 + ((float)g_vals[j]); // state update
s1_vals[j] = s1_vals[j]*s1_vals[j]; // update norm
break;
2023-03-09 18:10:19 +00:00
case LION:
2023-03-10 22:08:35 +00:00
s1_vals[j] = s1_vals[j]*beta2 + ((1.0f-beta2)*(float)g_vals[j]); // state update
2023-03-09 19:54:54 +00:00
break;
2022-10-27 11:11:29 +00:00
case RMSPROP:
2021-10-06 02:16:20 +00:00
s1_vals[j] = s1_vals[j]*beta1 + ((1.0f-beta1)*((float)g_vals[j])*((float)g_vals[j])); // state update
s1_vals[j] = __fdividef((float)g_vals[j],sqrtf(s1_vals[j])+eps); // update value
s1_vals[j] = s1_vals[j]*s1_vals[j]; // update norm
break;
2022-10-27 11:11:29 +00:00
case ADAGRAD:
2021-11-10 23:10:02 +00:00
s1_vals[j] = s1_vals[j] + ((float)g_vals[j])*((float)g_vals[j]); // state update
s1_vals[j] = __fdividef((float)g_vals[j],sqrtf(s1_vals[j])+eps); // update value
s1_vals[j] = s1_vals[j]*s1_vals[j]; // update norm
break;
2021-10-06 02:16:20 +00:00
}
}
# pragma unroll
for(unsigned int j = 1; j < NUM_VALS; j++)
s1_vals[0] += s1_vals[j];
__syncthreads();
s1_vals[0] = BlockReduce(temp_storage.reduce).Sum(s1_vals[0], valid_items);
if(threadIdx.x == 0)
atomicAdd(&unorm[0], s1_vals[0]);
__syncwarp();
}
}
template<typename T, int OPTIMIZER>
__launch_bounds__(TH, 1)
2022-10-27 11:11:29 +00:00
__global__ void kOptimizer32bit1State(T *g, T *p,
2021-10-06 02:16:20 +00:00
float *state1, float *unorm, const float max_unorm, const float param_norm,
2023-03-10 21:00:59 +00:00
const float beta1, const float beta2, const float eps, const float weight_decay,
2021-10-21 01:37:44 +00:00
const int step, const float lr, const float gnorm_scale, const bool skip_zeros, const int n)
2021-10-06 02:16:20 +00:00
{
const int n_full = ((TH*NUM_PER_THREAD)*(n/(TH*NUM_PER_THREAD))) + (n % (TH*NUM_PER_THREAD) == 0 ? 0 : (TH*NUM_PER_THREAD));
const int base_idx = (blockIdx.x * blockDim.x * NUM_PER_THREAD);
int valid_items = 0;
float update_scale = 0.0f;
if(max_unorm > 0.0f)
{
update_scale = max_unorm > 0.0f ? sqrtf(unorm[0]) : 1.0f;
if(update_scale > max_unorm*param_norm+eps){ update_scale = (max_unorm*param_norm+eps)/update_scale; }
else{ update_scale = 1.0f; }
}
else{ update_scale = 1.0f; }
T g_vals[NUM_PER_THREAD];
T p_vals[NUM_PER_THREAD];
float s1_vals[NUM_PER_THREAD];
typedef cub::BlockLoad<T, TH, NUM_PER_THREAD, cub::BLOCK_LOAD_WARP_TRANSPOSE> Load;
typedef cub::BlockStore<T, TH, NUM_PER_THREAD, cub::BLOCK_STORE_WARP_TRANSPOSE> Store;
typedef cub::BlockLoad<float, TH, NUM_PER_THREAD, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadFloat;
typedef cub::BlockStore<float, TH, NUM_PER_THREAD, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreFloat;
__shared__ union {
typename Load::TempStorage load;
typename Store::TempStorage store;
typename LoadFloat::TempStorage loadf;
typename StoreFloat::TempStorage storef;
} temp_storage;
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*TH*NUM_PER_THREAD)
{
valid_items = n - i >= (TH*NUM_PER_THREAD) ? (TH*NUM_PER_THREAD) : n - i;
__syncthreads();
Load(temp_storage.load).Load(&(g[i]), g_vals, valid_items);
__syncthreads();
LoadFloat(temp_storage.loadf).Load(&(state1[i]), s1_vals, valid_items);
__syncthreads();
Load(temp_storage.load).Load(&(p[i]), p_vals, valid_items);
# pragma unroll 4
for(unsigned int j = 0; j < NUM_PER_THREAD; j++)
{
g_vals[j] = gnorm_scale*((float)g_vals[j]);
if(weight_decay > 0.0f)
g_vals[j] = (float)g_vals[j] + (((float)p_vals[j])*weight_decay);
}
# pragma unroll 4
for(unsigned int j = 0; j < NUM_PER_THREAD; j++)
{
2021-10-22 00:22:43 +00:00
if(!skip_zeros || (skip_zeros && ((float)g_vals[j] != 0.0f)))
2021-10-21 02:15:47 +00:00
{
switch(OPTIMIZER)
{
2022-10-27 11:11:29 +00:00
case MOMENTUM:
2021-10-21 02:15:47 +00:00
if(step == 1)
s1_vals[j] = (float)g_vals[j];
else
s1_vals[j] = s1_vals[j]*beta1 + ((float)g_vals[j]);
p_vals[j] = ((float)p_vals[j]) + update_scale*(-lr*(s1_vals[j]));
break;
2023-03-09 19:54:54 +00:00
case LION:
p_vals[j] = ((float)p_vals[j]) - update_scale*(lr*sgn(((float)s1_vals[j])*beta1 + ((1.0f-beta1)*((float)g_vals[j]))));
2023-03-10 22:08:35 +00:00
s1_vals[j] = s1_vals[j]*beta2 + ((1.0f-beta2)*((float)g_vals[j]));
2023-03-09 19:54:54 +00:00
break;
2022-10-27 11:11:29 +00:00
case RMSPROP:
2021-10-21 02:15:47 +00:00
s1_vals[j] = s1_vals[j]*beta1 + ((1.0f-beta1)*((float)g_vals[j])*((float)g_vals[j]));
p_vals[j] = ((float)p_vals[j]) - update_scale*(lr*__fdividef((float)g_vals[j],sqrtf((float)s1_vals[j])+eps));
break;
2022-10-27 11:11:29 +00:00
case ADAGRAD:
2021-11-10 23:10:02 +00:00
s1_vals[j] = s1_vals[j] + ((float)g_vals[j])*((float)g_vals[j]);
p_vals[j] = ((float)p_vals[j]) - lr*__fdividef((float)g_vals[j],sqrtf((float)s1_vals[j])+eps);
break;
2021-10-21 02:15:47 +00:00
}
}
2021-10-06 02:16:20 +00:00
}
__syncthreads();
Store(temp_storage.store).Store(&(p[i]), p_vals, valid_items);
__syncthreads();
StoreFloat(temp_storage.storef).Store(&(state1[i]), s1_vals, valid_items);
}
}
#define NUM8BIT 16
#define NUM_THREADS 256
#define NUM_PER_BLOCK 4096
template<typename T, int OPTIMIZER>
__global__ void
__launch_bounds__(NUM_THREADS, 2)
kPreconditionOptimizerStatic8bit2State(T* p, T* __restrict__ const g, unsigned char*__restrict__ const state1, unsigned char* __restrict__ const state2,
float *unorm,
const float beta1, const float beta2,
const float eps, const int step,
float* __restrict__ const quantiles1, float* __restrict__ const quantiles2,
float* max1, float* max2, float* new_max1, float* new_max2,
const float gnorm_scale, const int n)
{
const int n_full = gridDim.x * NUM_PER_BLOCK;
const int base_idx = (blockIdx.x * blockDim.x * NUM_PER_THREAD);
int valid_items = n - (blockIdx.x*NUM_PER_BLOCK) > NUM_PER_BLOCK ? NUM_PER_BLOCK : n - (blockIdx.x*NUM_PER_BLOCK);
float g_val = 0.0f;
float local_max_s1 = -FLT_MAX;
float local_max_s2 = -FLT_MAX;
float local_unorm = 0.0f;
float s2_vals[NUM8BIT];
float s1_vals[NUM8BIT];
T g_vals[NUM8BIT];
unsigned char m_c1[NUM8BIT];
unsigned char r_c2[NUM8BIT];
typedef cub::BlockLoad<T, NUM_THREADS, NUM8BIT, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
typedef cub::BlockLoad<unsigned char, NUM_THREADS, NUM8BIT, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadUInt8;
typedef cub::BlockReduce<float, NUM_THREADS> BlockReduce;
__shared__ union {
typename LoadT::TempStorage loadh;
typename LoadUInt8::TempStorage loadc;
typename BlockReduce::TempStorage reduce;
} temp_storage;
__shared__ float smem_quantiles1[256];
__shared__ float smem_quantiles2[256];
if(threadIdx.x < 256)
{
smem_quantiles1[threadIdx.x] = quantiles1[threadIdx.x];
smem_quantiles2[threadIdx.x] = quantiles2[threadIdx.x];
}
__syncthreads();
for (unsigned int i = base_idx; i < n_full; i += NUM_THREADS*gridDim.x*NUM8BIT)
{
valid_items = n - i >= (TH*NUM_PER_THREAD) ? (TH*NUM_PER_THREAD) : n - i;
LoadT(temp_storage.loadh).Load(&(g[i]), g_vals, valid_items, (T)0.0f);
__syncthreads();
LoadUInt8(temp_storage.loadc).Load(&(state1[i]), m_c1, valid_items, 128);
__syncthreads();
LoadUInt8(temp_storage.loadc).Load(&(state2[i]), r_c2, valid_items, 128);
__syncthreads();
#pragma unroll 16
for(int j = 0; j < NUM8BIT; j++)
{
g_val = g_vals[j];
g_val *= gnorm_scale;
s1_vals[j] = smem_quantiles1[m_c1[j]]*max1[0]*beta1;
s1_vals[j] += (1.0f-beta1)*g_val;
local_max_s1 = fmaxf(local_max_s1, fabsf(s1_vals[j]));
}
#pragma unroll 16
for(int j = 0; j < NUM8BIT; j++)
{
g_val = g_vals[j];
g_val *= gnorm_scale;
s2_vals[j] = smem_quantiles2[r_c2[j]]*max2[0]*beta2;
s2_vals[j] += (1.0f-beta2)*g_val*g_val;
local_max_s2 = fmaxf(local_max_s2, fabsf(s2_vals[j]));
}
if(unorm != NULL)
{
#pragma unroll 16
for(int j = 0; j < NUM8BIT; j++)
{
float correction1 = __fdividef(1.0f, 1.0f - powf(beta1, step));
float correction2 = __fdividef(1.0f, 1.0f - powf(beta2, step));
s1_vals[j] *= correction1;
s2_vals[j] *= correction2;
float update_val = s1_vals[j]/(sqrtf(s2_vals[j])+eps); // update
local_unorm += update_val*update_val;
}
}
}
__syncthreads();
local_max_s1 = BlockReduce(temp_storage.reduce).Reduce(local_max_s1, cub::Max(), valid_items);
__syncthreads();
local_max_s2 = BlockReduce(temp_storage.reduce).Reduce(local_max_s2, cub::Max(), valid_items);
if(unorm != NULL)
{
__syncthreads();
local_unorm = BlockReduce(temp_storage.reduce).Reduce(local_unorm, cub::Sum(), valid_items);
}
if(threadIdx.x == 0)
{
atomicMax(&new_max1[0], local_max_s1);
atomicMax(&new_max2[0], local_max_s2);
if(unorm != NULL){ atomicAdd(&unorm[0], local_unorm); }
}
}
#define NUM_PER_THREAD2 4
#define NUM_THREADS2 1024
#define NUM_PER_BLOCK2 4096
template<typename T, int OPTIMIZER>
__global__ void
__launch_bounds__(NUM_THREADS2, 1)
kOptimizerStatic8bit2State(T* p, T* const g, unsigned char* state1, unsigned char* state2,
const float *unorm, const float max_unorm, const float param_norm, \
const float beta1, const float beta2,
const float eps, const int step, const float lr,
float* __restrict__ const quantiles1, float* __restrict__ const quantiles2,
float* max1, float* max2, float* new_max1, float* new_max2,
float weight_decay,
const float gnorm_scale, const int n)
{
const int n_full = (blockDim.x * gridDim.x)*NUM_PER_THREAD2;
const int base_idx = (blockIdx.x * blockDim.x * NUM_PER_THREAD2);
int valid_items = 0;
float g_val = 0.0f;
float s1_vals[NUM_PER_THREAD2];
float s2_vals[NUM_PER_THREAD2];
const float correction1 = 1.0f - powf(beta1, step);
const float correction2 = sqrtf(1.0f - powf(beta2, step));
const float step_size = -lr*correction2/correction1;
//const float step_size = -lr*correction2/correction1;
float new_max_val1 = 1.0f/new_max1[0];
float new_max_val2 = 1.0f/new_max2[0];
float update_scale = 1.0f;
if(max_unorm > 0.0f)
{
update_scale = max_unorm > 0.0f ? sqrtf(unorm[0]) : 1.0f;
if(update_scale > max_unorm*param_norm){ update_scale = (max_unorm*param_norm)/update_scale; }
else{ update_scale = 1.0f; }
}
else{ update_scale = 1.0f; }
unsigned char c1s[NUM_PER_THREAD2];
unsigned char c2s[NUM_PER_THREAD2];
T p_vals[NUM_PER_THREAD2];
T g_vals[NUM_PER_THREAD2];
typedef cub::BlockLoad<T, NUM_THREADS2, NUM_PER_THREAD2, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
typedef cub::BlockLoad<unsigned char, NUM_THREADS2, NUM_PER_THREAD2, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadChar;
typedef cub::BlockStore<unsigned char, NUM_THREADS2, NUM_PER_THREAD2, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreChar;
typedef cub::BlockStore<T, NUM_THREADS2, NUM_PER_THREAD2, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreT;
__shared__ float smem_quantiles1[256];
__shared__ float smem_quantiles2[256];
__shared__ union {
typename LoadT::TempStorage loadh;
typename LoadChar::TempStorage loadc;
typename StoreChar::TempStorage storec;
typename StoreT::TempStorage storeh;
} temp_storage;
if(threadIdx.x < 512)
{
if(threadIdx.x < 256)
smem_quantiles1[threadIdx.x] = quantiles1[threadIdx.x];
else
smem_quantiles2[threadIdx.x-256] = quantiles2[threadIdx.x-256];
}
__syncthreads();
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*NUM_THREADS2*NUM_PER_THREAD2)
{
valid_items = n - i >= (TH*NUM_PER_THREAD) ? (TH*NUM_PER_THREAD) : n - i;
LoadT(temp_storage.loadh).Load(&(g[i]), g_vals, valid_items, (T)0.0f);
__syncthreads();
LoadChar(temp_storage.loadc).Load(&(state1[i]), c1s, valid_items, 128);
__syncthreads();
LoadChar(temp_storage.loadc).Load(&(state2[i]), c2s, valid_items, 0);
__syncthreads();
LoadT(temp_storage.loadh).Load(&(p[i]), p_vals, valid_items);
if((i + (threadIdx.x*NUM_PER_THREAD2) + NUM_PER_THREAD2) > n){ continue; }
# pragma unroll 4
for(unsigned int j = 0; j < NUM_PER_THREAD2; j++)
{
g_val = float(g_vals[j]);
g_val *= gnorm_scale;
s1_vals[j] = smem_quantiles1[c1s[j]];
s1_vals[j] = s1_vals[j]*max1[0];
s1_vals[j] = (s1_vals[j]*beta1) + (((1.0f-beta1)*g_val));
c1s[j] = dQuantize<0>(smem_quantiles1, 0.0f, s1_vals[j]*new_max_val1);
// make sure state1 term has still the same sign after quantization
// (not needed for state2 term which has only positive values)
if(signbit(smem_quantiles1[c1s[j]]) != signbit(s1_vals[j]))
{
if(s1_vals[j] > 0.0f)
c1s[j] += 1;
else
c1s[j] -= 1;
}
s2_vals[j] = smem_quantiles2[c2s[j]];
s2_vals[j] = s2_vals[j]*max2[0];
s2_vals[j] = (s2_vals[j]*beta2) + (((1.0f-beta2)*g_val*g_val));
c2s[j] = dQuantize<0>(smem_quantiles2, 0.0f, s2_vals[j]*new_max_val2);
}
# pragma unroll 4
for(unsigned int j = 0; j < NUM_PER_THREAD2; j++)
{
p_vals[j] = (T)(((float)p_vals[j]) + ((update_scale*step_size*(s1_vals[j]/(sqrtf(s2_vals[j])+(correction2*eps))))));
if(weight_decay > 0.0f)
p_vals[j] = update_scale*((float)p_vals[j])*(1.0f-(lr*weight_decay));
}
StoreT(temp_storage.storeh).Store(&(p[i]), p_vals, valid_items);
__syncthreads();
StoreChar(temp_storage.storec).Store(&(state1[i]), c1s, valid_items);
__syncthreads();
StoreChar(temp_storage.storec).Store(&(state2[i]), c2s, valid_items);
__syncthreads();
}
}
template<typename T, int OPTIMIZER>
__global__ void
__launch_bounds__(NUM_THREADS, 2)
2022-10-27 11:11:29 +00:00
kPreconditionOptimizerStatic8bit1State(T* p, T* __restrict__ const g, unsigned char*__restrict__ const state1,
2021-10-06 02:16:20 +00:00
float *unorm,
2023-03-10 21:00:59 +00:00
const float beta1, const float beta2,
2021-10-06 02:16:20 +00:00
const float eps, const int step,
2022-10-27 11:11:29 +00:00
float* __restrict__ const quantiles1,
float* max1, float* new_max1,
2021-10-06 02:16:20 +00:00
const float weight_decay,
const float gnorm_scale, const int n)
{
const int n_full = gridDim.x * NUM_PER_BLOCK;
const int base_idx = (blockIdx.x * blockDim.x * NUM_PER_THREAD);
int valid_items = n - (blockIdx.x*NUM_PER_BLOCK) > NUM_PER_BLOCK ? NUM_PER_BLOCK : n - (blockIdx.x*NUM_PER_BLOCK);
float g_val = 0.0f;
float local_max_s1 = -FLT_MAX;
float local_unorm = 0.0f;
float s1_vals[NUM8BIT];
T g_vals[NUM8BIT];
unsigned char m_c1[NUM8BIT];
typedef cub::BlockLoad<T, NUM_THREADS, NUM8BIT, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
typedef cub::BlockLoad<unsigned char, NUM_THREADS, NUM8BIT, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadUInt8;
typedef cub::BlockReduce<float, NUM_THREADS> BlockReduce;
__shared__ union {
typename LoadT::TempStorage loadh;
typename LoadUInt8::TempStorage loadc;
typename BlockReduce::TempStorage reduce;
} temp_storage;
__shared__ float smem_quantiles1[256];
if(threadIdx.x < 256)
smem_quantiles1[threadIdx.x] = quantiles1[threadIdx.x];
__syncthreads();
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*NUM_THREADS*NUM8BIT)
{
valid_items = n - i >= (TH*NUM_PER_THREAD) ? (TH*NUM_PER_THREAD) : n - i;
__syncthreads();
LoadT(temp_storage.loadh).Load(&(g[i]), g_vals, valid_items, (T)0.0f);
__syncthreads();
LoadUInt8(temp_storage.loadc).Load(&(state1[i]), m_c1, valid_items, 128);
#pragma unroll 16
for(int j = 0; j < NUM8BIT; j++)
{
g_val = g_vals[j];
g_val *= gnorm_scale;
s1_vals[j] = smem_quantiles1[m_c1[j]]*max1[0];
switch(OPTIMIZER)
{
2022-10-27 11:11:29 +00:00
case MOMENTUM:
2021-10-06 02:16:20 +00:00
if(step == 1)
s1_vals[j] = (float)g_vals[j];
else
s1_vals[j] = s1_vals[j]*beta1 + ((float)g_vals[j]);
if(unorm != NULL)
local_unorm += s1_vals[j]*s1_vals[j];
break;
2023-03-09 18:10:19 +00:00
case LION:
2023-03-10 21:00:59 +00:00
s1_vals[j] = s1_vals[j]*beta2 + ((1.0f-beta2)*g_val);
2023-03-09 19:54:54 +00:00
break;
2022-10-27 11:11:29 +00:00
case RMSPROP:
2021-10-06 02:16:20 +00:00
s1_vals[j] = s1_vals[j]*beta1 + ((1.0f-beta1)*(g_val*g_val));
break;
}
local_max_s1 = fmaxf(local_max_s1, fabsf(s1_vals[j]));
}
}
__syncthreads();
local_max_s1 = BlockReduce(temp_storage.reduce).Reduce(local_max_s1, cub::Max(), valid_items);
if(threadIdx.x == 0){ atomicMax(&new_max1[0], local_max_s1); }
if(unorm != NULL)
{
__syncthreads();
local_unorm = BlockReduce(temp_storage.reduce).Reduce(local_unorm, cub::Sum(), valid_items);
if(threadIdx.x == 0){ atomicAdd(&unorm[0], local_unorm); }
}
}
template<typename T, int OPTIMIZER>
__global__ void
2023-04-11 15:37:02 +00:00
__launch_bounds__(1024, 1)
2021-10-06 02:16:20 +00:00
kOptimizerStatic8bit1State(T* p, T* const g, unsigned char* state1,
const float *unorm, const float max_unorm, const float param_norm,
2023-03-10 21:00:59 +00:00
const float beta1, const float beta2,
2021-10-06 02:16:20 +00:00
const float eps, const int step, const float lr,
2022-10-27 11:11:29 +00:00
float* __restrict__ const quantiles1,
float* max1, float* new_max1,
2021-10-06 02:16:20 +00:00
float weight_decay,
const float gnorm_scale, const int n)
{
const int n_full = (blockDim.x * gridDim.x)*NUM_PER_THREAD2;
const int base_idx = (blockIdx.x * blockDim.x * NUM_PER_THREAD2);
int valid_items = 0;
float g_val = 0.0f;
float s1_vals[NUM_PER_THREAD2];
float new_max_val1 = 1.0f/new_max1[0];
float update_scale = 1.0f;
if(max_unorm > 0.0f)
{
update_scale = max_unorm > 0.0f ? sqrtf(unorm[0]) : 1.0f;
if(update_scale > max_unorm*param_norm){ update_scale = (max_unorm*param_norm)/update_scale; }
else{ update_scale = 1.0f; }
}
else{ update_scale = 1.0f; }
unsigned char c1s[NUM_PER_THREAD2];
T p_vals[NUM_PER_THREAD2];
T g_vals[NUM_PER_THREAD2];
typedef cub::BlockLoad<T, NUM_THREADS2, NUM_PER_THREAD2, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
typedef cub::BlockLoad<unsigned char, NUM_THREADS2, NUM_PER_THREAD2, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadChar;
typedef cub::BlockStore<unsigned char, NUM_THREADS2, NUM_PER_THREAD2, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreChar;
typedef cub::BlockStore<T, NUM_THREADS2, NUM_PER_THREAD2, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreT;
__shared__ float smem_quantiles1[256];
__shared__ union {
typename LoadT::TempStorage loadh;
typename LoadChar::TempStorage loadc;
typename StoreChar::TempStorage storec;
typename StoreT::TempStorage storeh;
} temp_storage;
if(threadIdx.x < 256)
smem_quantiles1[threadIdx.x] = quantiles1[threadIdx.x];
__syncthreads();
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*NUM_THREADS2*NUM_PER_THREAD2)
{
valid_items = n - i >= (TH*NUM_PER_THREAD) ? (TH*NUM_PER_THREAD) : n - i;
LoadT(temp_storage.loadh).Load(&(g[i]), g_vals, valid_items, (T)0.0f);
__syncthreads();
LoadChar(temp_storage.loadc).Load(&(state1[i]), c1s, valid_items, 128);
__syncthreads();
LoadT(temp_storage.loadh).Load(&(p[i]), p_vals, valid_items);
if((i + (threadIdx.x*NUM_PER_THREAD2) + NUM_PER_THREAD2) > n){ continue; }
# pragma unroll 4
for(unsigned int j = 0; j < NUM_PER_THREAD2; j++)
{
g_val = float(g_vals[j]);
g_val *= gnorm_scale;
2023-03-09 22:03:07 +00:00
if(weight_decay > 0.0f) {
switch(OPTIMIZER) {
case MOMENTUM:
case RMSPROP:
g_val += ((float)p_vals[j])*weight_decay;
break;
case LION:
p_vals[j] = ((float)p_vals[j])*(1.0f-lr*weight_decay);
break;
}
}
2021-10-06 02:16:20 +00:00
s1_vals[j] = smem_quantiles1[c1s[j]]*max1[0];
switch(OPTIMIZER)
{
2022-10-27 11:11:29 +00:00
case MOMENTUM:
2021-10-06 02:16:20 +00:00
if(step == 1)
s1_vals[j] = g_vals[j];
else
s1_vals[j] = s1_vals[j]*beta1 + ((float)g_vals[j]);
p_vals[j] = ((float)p_vals[j]) + (-lr*update_scale*(s1_vals[j]));
break;
2023-03-09 18:10:19 +00:00
case LION:
2023-03-09 19:54:54 +00:00
p_vals[j] = ((float)p_vals[j]) - (lr*sgn(((float)s1_vals[j])*beta1 + ((1.0f-beta1)*((float)g_val))));
2023-03-10 21:00:59 +00:00
s1_vals[j] = s1_vals[j]*beta2 + ((1.0f-beta2)*g_val);
2023-03-09 19:54:54 +00:00
break;
2022-10-27 11:11:29 +00:00
case RMSPROP:
2021-10-06 02:16:20 +00:00
s1_vals[j] = s1_vals[j]*beta1 + ((1.0f-beta1)*(g_val*g_val));
p_vals[j] = ((float)p_vals[j]) - (lr*__fdividef(g_val,sqrtf(s1_vals[j])+eps));
break;
}
c1s[j] = dQuantize<0>(smem_quantiles1, 0.0f, s1_vals[j]*new_max_val1);
// make sure state1 term has still the same sign after quantization
if(signbit(smem_quantiles1[c1s[j]]) != signbit(s1_vals[j]))
{
if(s1_vals[j] > 0.0f)
c1s[j] += 1;
else
c1s[j] -= 1;
}
}
StoreT(temp_storage.storeh).Store(&(p[i]), p_vals, valid_items);
__syncthreads();
StoreChar(temp_storage.storec).Store(&(state1[i]), c1s, valid_items);
__syncthreads();
}
}
template<typename T, int BLOCK_SIZE, int NUM_VALS>
__global__ void kPercentileClipping(T * __restrict__ g, float *gnorm_vec, int step, const int n)
{
const int n_full = (BLOCK_SIZE*(n/BLOCK_SIZE)) + (n % BLOCK_SIZE == 0 ? 0 : BLOCK_SIZE);
int valid_items = 0;
typedef cub::BlockReduce<float, BLOCK_SIZE/NUM_VALS> BlockReduce;
typedef cub::BlockLoad<T, BLOCK_SIZE/NUM_VALS, NUM_VALS, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
__shared__ typename BlockReduce::TempStorage reduce;
__shared__ typename LoadT::TempStorage loadT;
T vals[NUM_VALS];
float local_sum = 0.0f;
for (unsigned int i = (blockIdx.x * BLOCK_SIZE); i < n_full; i += gridDim.x*BLOCK_SIZE)
{
valid_items = n - i > BLOCK_SIZE ? BLOCK_SIZE : n - i;
local_sum = 0.0f;
__syncthreads();
LoadT(loadT).Load(&(g[i]), vals, valid_items, (T)0.0f);
#pragma unroll NUM_VALS
for(int j = 0; j < NUM_VALS; j++)
local_sum += ((float)vals[j])*((float)vals[j]);
local_sum = BlockReduce(reduce).Sum(local_sum, valid_items);
if(threadIdx.x == 0)
{
if(step == 1)
{
// initialize with the same norm for all positions
//#pragma unroll 10
for(int j = 0; j < 100; j++)
atomicAdd(&gnorm_vec[j], local_sum);
}
else
atomicAdd(&gnorm_vec[step % 100], local_sum);
}
}
}
#define LANES 2
#define QUAD 3
template<typename T, int OPTIMIZER, int BLOCK_SIZE, int N_PER_TH>
__launch_bounds__(256, 3)
__global__ void
kOptimizerStatic8bit2StateBlockwise(T* p, T* __restrict__ const g, unsigned char* state1, unsigned char* state2,
const float beta1, const float beta2,
const float eps, const int step, const float lr,
float* __restrict__ const quantiles1, float* __restrict__ const quantiles2,
2022-10-27 11:11:29 +00:00
float* absmax1, float* absmax2,
2021-10-06 02:16:20 +00:00
float weight_decay,
2021-10-21 01:37:44 +00:00
const float gnorm_scale, const bool skip_zeros, const int n)
2021-10-06 02:16:20 +00:00
{
//const int n_full = n + (n%BLOCK_SIZE);
const int n_full = gridDim.x * BLOCK_SIZE;
const int base_idx = (blockIdx.x * BLOCK_SIZE);
int valid_items = 0;
float g_val = 0.0f;
float s1_vals[N_PER_TH];
float s2_vals[N_PER_TH];
// 2-5%
const float correction1 = 1.0f - __powf(beta1, step);
const float correction2 = sqrtf(1.0f -__powf(beta2, step));
const float step_size = __fdividef(-lr*correction2,correction1);
const int lane_id = threadIdx.x % LANES;
float new_local_abs_max1 = -FLT_MAX;
float new_local_abs_max2 = -FLT_MAX;
float quadrants1[QUAD];
float quadrants2[QUAD];
unsigned char c1s[N_PER_TH];
unsigned char c2s[N_PER_TH];
T g_vals[N_PER_TH];
2023-04-07 16:59:21 +00:00
T p_vals[N_PER_TH];
2021-10-06 02:16:20 +00:00
typedef cub::BlockLoad<T, BLOCK_SIZE/N_PER_TH, N_PER_TH, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
typedef cub::BlockLoad<unsigned char, BLOCK_SIZE/N_PER_TH, N_PER_TH, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadChar;
typedef cub::BlockStore<unsigned char, BLOCK_SIZE/N_PER_TH, N_PER_TH, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreChar;
typedef cub::BlockStore<T, BLOCK_SIZE/N_PER_TH, N_PER_TH, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreT;
__shared__ float smem_quantiles1[LANES][257];
__shared__ float smem_quantiles2[LANES][257];
typedef cub::BlockReduce<float, BLOCK_SIZE/N_PER_TH> BlockReduce1;
typedef cub::BlockReduce<float, BLOCK_SIZE/N_PER_TH> BlockReduce2;
__shared__ typename BlockReduce1::TempStorage reduce1;
__shared__ typename BlockReduce2::TempStorage reduce2;
__shared__ float smem_exchange1[1];
__shared__ float smem_exchange2[1];
__shared__ union {
typename LoadT::TempStorage loadh;
typename LoadChar::TempStorage loadc;
typename StoreChar::TempStorage storec;
typename StoreT::TempStorage storeh;
} temp_storage;
// init: 0.2 -> 0.23
// 0.23 -> 0.23
smem_quantiles1[0][threadIdx.x] = quantiles1[threadIdx.x];
smem_quantiles2[0][threadIdx.x] = quantiles2[threadIdx.x];
# pragma unroll
for(unsigned int j = 1; j < LANES; j++)
{
smem_quantiles1[j][threadIdx.x] = smem_quantiles1[0][threadIdx.x];
smem_quantiles2[j][threadIdx.x] = smem_quantiles2[0][threadIdx.x];
}
__syncthreads();
#pragma unroll
for(int k = 0; k < QUAD; k++)
{
quadrants1[k] = smem_quantiles1[lane_id][(k*256/(QUAD+1)) + (256/(QUAD+1)-1)];
quadrants2[k] = smem_quantiles2[lane_id][(k*256/(QUAD+1)) + (256/(QUAD+1)-1)];
}
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*BLOCK_SIZE)
{
// loads: 0.23 -> 0.85/1.44
valid_items = n - i >= BLOCK_SIZE ? BLOCK_SIZE : n - i;
__syncthreads();
LoadT(temp_storage.loadh).Load(&(g[i]), g_vals, valid_items, (T)0.0f);
__syncthreads();
LoadChar(temp_storage.loadc).Load(&(state1[i]), c1s, valid_items, 128);
__syncthreads();
LoadChar(temp_storage.loadc).Load(&(state2[i]), c2s, valid_items, 0);
new_local_abs_max1 = -FLT_MAX;
new_local_abs_max2 = -FLT_MAX;
// update: 2.48/1.57 -> 2.51/1.60
# pragma unroll N_PER_TH
for(unsigned int j = 0; j < N_PER_TH; j++)
{
2023-04-07 16:59:21 +00:00
if(!isnan((float)g_vals[j]) && !isinf((float)g_vals[j]))
2021-10-21 02:15:47 +00:00
{
s2_vals[j] = smem_quantiles2[lane_id][c2s[j]]*absmax2[i/BLOCK_SIZE];
2023-04-07 16:59:21 +00:00
g_val = g_vals[j];
//float ratio = (g_val*g_val)/fmaxf(s2_vals[j], eps*eps);
//g_val = ratio > 2.0f ? 2.0f*g_val/ratio : g_val;
g_val *= gnorm_scale;
2021-10-21 02:15:47 +00:00
s2_vals[j] = (s2_vals[j]*beta2) + (((1.0f-beta2)*g_val*g_val));
2023-04-07 16:59:21 +00:00
s1_vals[j] = smem_quantiles1[lane_id][c1s[j]]*absmax1[i/BLOCK_SIZE];
s1_vals[j] = (s1_vals[j]*beta1) + (((1.0f-beta1)*g_val));
2021-10-21 02:15:47 +00:00
}
2023-04-07 16:59:21 +00:00
else
{
s1_vals[j] = 0.0f;
s2_vals[j] = 0.0f;
}
2021-10-06 02:16:20 +00:00
new_local_abs_max1 = fmaxf(new_local_abs_max1, fabsf(s1_vals[j]));
new_local_abs_max2 = fmaxf(new_local_abs_max2, fabsf(s2_vals[j]));
}
// reduce: 2.51/1.60 -> 2.67/1.69
new_local_abs_max1 = BlockReduce1(reduce1).Reduce(new_local_abs_max1, cub::Max());
new_local_abs_max2 = BlockReduce2(reduce2).Reduce(new_local_abs_max2, cub::Max());
if(threadIdx.x == 0)
{
smem_exchange1[0] = new_local_abs_max1;
smem_exchange2[0] = new_local_abs_max2;
}
__syncthreads();
if(threadIdx.x == 0)
{
absmax1[i/BLOCK_SIZE] = new_local_abs_max1;
absmax2[i/BLOCK_SIZE] = new_local_abs_max2;
}
else
{
new_local_abs_max1 = smem_exchange1[0];
new_local_abs_max2 = smem_exchange2[0];
}
__syncthreads();
2023-04-07 16:59:21 +00:00
LoadT(temp_storage.loadh).Load(&(p[i]), p_vals, valid_items, (T)0.0f);
2021-10-06 02:16:20 +00:00
// reduce: 2.67/1.69 -> 2.67/1.70
# pragma unroll N_PER_TH
for(unsigned int j = 0; j < N_PER_TH; j++)
{
2023-04-07 16:59:21 +00:00
//if(!skip_zeros || (skip_zeros && ((float)g_vals[j] != 0.0f)))
if(!isnan((float)g_vals[j]) && !isinf((float)g_vals[j]))
2021-10-21 02:15:47 +00:00
{
2023-04-07 16:59:21 +00:00
p_vals[j] = (T)(((float)p_vals[j]) + ((step_size*(__fdividef(s1_vals[j],(sqrtf(s2_vals[j])+(correction2*eps)))))));
2021-10-21 02:15:47 +00:00
if(weight_decay > 0.0f)
2023-04-07 16:59:21 +00:00
p_vals[j] = ((float)p_vals[j])*(1.0f-(lr*weight_decay));
2021-10-21 02:15:47 +00:00
}
2021-10-06 02:16:20 +00:00
}
// store: 0.85/1.44 -> 2.48/1.57
__syncthreads();
2023-04-07 16:59:21 +00:00
StoreT(temp_storage.storeh).Store(&(p[i]), p_vals, valid_items);
2021-10-06 02:16:20 +00:00
// quantizaztion: 2.67/1.70 -> 3.4/3.3
2022-10-27 11:11:29 +00:00
# pragma unroll N_PER_TH
2021-10-06 02:16:20 +00:00
for(unsigned int j = 0; j < N_PER_TH; j++)
{
c1s[j] = quantize_2D<1>(quadrants1, smem_quantiles1[lane_id], __fdividef(s1_vals[j],new_local_abs_max1));
c2s[j] = quantize_2D<0>(quadrants2, smem_quantiles2[lane_id], __fdividef(s2_vals[j],new_local_abs_max2));
// make sure state1 term has still the same sign after quantization
// (not needed for state2 term which has only positive values)
if(signbit(smem_quantiles1[lane_id][c1s[j]]) != signbit(s1_vals[j]))
{
if(s1_vals[j] > 0.0f)
c1s[j] += 1;
else
c1s[j] -= 1;
}
}
__syncthreads();
StoreChar(temp_storage.storec).Store(&(state1[i]), c1s, valid_items);
__syncthreads();
StoreChar(temp_storage.storec).Store(&(state2[i]), c2s, valid_items);
}
}
#define LANES 2
#define QUAD 3
template<typename T, int OPTIMIZER, int BLOCK_SIZE, int N_PER_TH>
__launch_bounds__(256, 3)
__global__ void
kOptimizerStatic8bit1StateBlockwise(T* p, T* __restrict__ const g, unsigned char* state1,
const float beta1, const float beta2,
const float eps, const int step, const float lr,
float* __restrict__ const quantiles1,
float* absmax1,
float weight_decay,
2021-10-21 01:37:44 +00:00
const float gnorm_scale, const bool skip_zeros, const int n)
2021-10-06 02:16:20 +00:00
{
//const int n_full = n + (n%BLOCK_SIZE);
const int n_full = gridDim.x * BLOCK_SIZE;
const int base_idx = (blockIdx.x * BLOCK_SIZE);
int valid_items = 0;
float g_val = 0.0f;
float s1_vals[N_PER_TH];
// 2-5%
const int lane_id = threadIdx.x % LANES;
float new_local_abs_max1 = -FLT_MAX;
float quadrants1[QUAD];
unsigned char c1s[N_PER_TH];
T g_vals[N_PER_TH];
T p_vals[N_PER_TH];
typedef cub::BlockLoad<T, BLOCK_SIZE/N_PER_TH, N_PER_TH, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadT;
typedef cub::BlockLoad<unsigned char, BLOCK_SIZE/N_PER_TH, N_PER_TH, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadChar;
typedef cub::BlockStore<unsigned char, BLOCK_SIZE/N_PER_TH, N_PER_TH, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreChar;
typedef cub::BlockStore<T, BLOCK_SIZE/N_PER_TH, N_PER_TH, cub::BLOCK_STORE_WARP_TRANSPOSE> StoreT;
__shared__ float smem_quantiles1[LANES][257];
typedef cub::BlockReduce<float, BLOCK_SIZE/N_PER_TH> BlockReduce1;
__shared__ typename BlockReduce1::TempStorage reduce1;
__shared__ float smem_exchange1[1];
__shared__ union {
typename LoadT::TempStorage loadh;
typename LoadChar::TempStorage loadc;
typename StoreChar::TempStorage storec;
typename StoreT::TempStorage storeh;
} temp_storage;
// init: 0.2 -> 0.23
// 0.23 -> 0.23
smem_quantiles1[0][threadIdx.x] = quantiles1[threadIdx.x];
# pragma unroll
for(unsigned int j = 1; j < LANES; j++)
smem_quantiles1[j][threadIdx.x] = smem_quantiles1[0][threadIdx.x];
__syncthreads();
#pragma unroll
for(int k = 0; k < QUAD; k++)
quadrants1[k] = smem_quantiles1[lane_id][(k*256/(QUAD+1)) + (256/(QUAD+1)-1)];
for (unsigned int i = base_idx; i < n_full; i += gridDim.x*BLOCK_SIZE)
{
// loads: 0.23 -> 0.85/1.44
valid_items = n - i >= BLOCK_SIZE ? BLOCK_SIZE : n - i;
__syncthreads();
LoadT(temp_storage.loadh).Load(&(g[i]), g_vals, valid_items, (T)0.0f);
__syncthreads();
LoadChar(temp_storage.loadc).Load(&(state1[i]), c1s, valid_items, 128);
__syncthreads();
LoadT(temp_storage.loadh).Load(&(p[i]), p_vals, valid_items, (T)0.0f);
new_local_abs_max1 = -FLT_MAX;
// update: 2.48/1.57 -> 2.51/1.60
# pragma unroll N_PER_TH
for(unsigned int j = 0; j < N_PER_TH; j++)
{
g_val = float(g_vals[j]);
g_val *= gnorm_scale;
2023-03-22 16:33:47 +00:00
if(!skip_zeros || (skip_zeros && ((float)g_vals[j] != 0.0f)))
{
if(weight_decay > 0.0f) {
switch(OPTIMIZER) {
2023-03-09 22:03:07 +00:00
case MOMENTUM:
2023-03-09 22:05:45 +00:00
case ADAGRAD:
2023-03-09 22:03:07 +00:00
case RMSPROP:
g_val += ((float)p_vals[j])*weight_decay;
break;
case LION:
p_vals[j] = ((float)p_vals[j])*(1.0f-lr*weight_decay);
break;
}
}
2021-10-21 02:15:47 +00:00
s1_vals[j] = smem_quantiles1[lane_id][c1s[j]]*absmax1[i/BLOCK_SIZE];
switch(OPTIMIZER)
{
2022-10-27 11:11:29 +00:00
case MOMENTUM:
2021-10-21 02:15:47 +00:00
if(step == 1)
s1_vals[j] = g_val;
else
s1_vals[j] = (s1_vals[j]*beta1) + g_val;
break;
2023-03-09 19:54:54 +00:00
case LION:
2023-03-22 16:56:50 +00:00
// here, using gvals[j] to store the gradient smoothed by beta1 for the following parameter update, before the momentum is updated by beta2
2023-03-22 16:27:13 +00:00
g_vals[j] = lr*sgn(((float)s1_vals[j])*beta1 + ((1.0f-beta1)*g_val));
2023-03-10 20:50:14 +00:00
s1_vals[j] = s1_vals[j]*beta2 + ((1.0f-beta2)*g_val);
2023-03-09 19:54:54 +00:00
break;
2022-10-27 11:11:29 +00:00
case RMSPROP:
2021-10-21 02:15:47 +00:00
s1_vals[j] = s1_vals[j]*beta1 + ((1.0f-beta1)*(g_val*g_val));
break;
2022-10-27 11:11:29 +00:00
case ADAGRAD:
2021-11-10 23:10:02 +00:00
s1_vals[j] = s1_vals[j] + (g_val*g_val);
break;
2021-10-21 02:15:47 +00:00
}
}
2021-10-06 02:16:20 +00:00
new_local_abs_max1 = fmaxf(new_local_abs_max1, fabsf(s1_vals[j]));
}
// reduce: 2.51/1.60 -> 2.67/1.69
new_local_abs_max1 = BlockReduce1(reduce1).Reduce(new_local_abs_max1, cub::Max());
if(threadIdx.x == 0)
smem_exchange1[0] = new_local_abs_max1;
__syncthreads();
if(threadIdx.x == 0)
absmax1[i/BLOCK_SIZE] = new_local_abs_max1;
else
new_local_abs_max1 = smem_exchange1[0];
// reduce: 2.67/1.69 -> 2.67/1.70
# pragma unroll N_PER_TH
for(unsigned int j = 0; j < N_PER_TH; j++)
{
2021-10-22 00:22:43 +00:00
if(!skip_zeros || (skip_zeros && ((float)g_vals[j] != 0.0f)))
2021-10-21 02:15:47 +00:00
{
switch(OPTIMIZER)
{
2022-10-27 11:11:29 +00:00
case MOMENTUM:
2021-10-21 02:15:47 +00:00
p_vals[j] = ((float)p_vals[j]) - lr*(s1_vals[j]);
break;
2023-03-09 19:54:54 +00:00
case LION:
2023-03-22 14:52:59 +00:00
p_vals[j] = ((float)p_vals[j]) - ((float)g_vals[j]);
2023-03-09 19:54:54 +00:00
break;
2022-10-27 11:11:29 +00:00
case RMSPROP:
2021-10-21 02:15:47 +00:00
g_val = g_vals[j];
p_vals[j] = ((float)p_vals[j]) - lr*(__fdividef(g_val, sqrtf(s1_vals[j])+eps));
break;
2022-10-27 11:11:29 +00:00
case ADAGRAD:
2021-11-10 23:10:02 +00:00
g_val = g_vals[j];
p_vals[j] = ((float)p_vals[j]) - lr*(__fdividef(g_val, sqrtf(s1_vals[j])+eps));
break;
2021-10-21 02:15:47 +00:00
}
}
2021-10-06 02:16:20 +00:00
}
// store: 0.85/1.44 -> 2.48/1.57
__syncthreads();
StoreT(temp_storage.storeh).Store(&(p[i]), p_vals, valid_items);
// quantizaztion: 2.67/1.70 -> 3.4/3.3
2022-10-27 11:11:29 +00:00
# pragma unroll N_PER_TH
2021-10-06 02:16:20 +00:00
for(unsigned int j = 0; j < N_PER_TH; j++)
{
c1s[j] = quantize_2D<1>(quadrants1, smem_quantiles1[lane_id], __fdividef(s1_vals[j],new_local_abs_max1));
// make sure state1 term has still the same sign after quantization
// (not needed for state2 term which has only positive values)
if(signbit(smem_quantiles1[lane_id][c1s[j]]) != signbit(s1_vals[j]))
{
if(s1_vals[j] > 0.0f)
c1s[j] += 1;
else
c1s[j] -= 1;
}
}
__syncthreads();
StoreChar(temp_storage.storec).Store(&(state1[i]), c1s, valid_items);
}
}
2022-07-22 21:41:05 +00:00
template<typename T, int THREADS, int ITEMS_PER_THREAD, int TILE_ROWS, int TILE_COLS, int SPARSE_DECOMP> __global__ void kgetColRowStats(T * __restrict__ A, float *rowStats, float *colStats, int * nnz_count_row, float nnz_threshold, int rows, int cols, int tiledRows, int tiledCols)
{
// 0. reset stats to -FLT_MAX
// 1. load row-by-row ITEMS_PER_THREAD (TILE_SIZE==THREADS*ITEMS_PER_THREAD)
// 2. compute col max (per thread); store in smem due to register pressure
// 3. compute row max (per block); store in smem to accumulate full global mem transation
// 4. store data via atomicMax
// each block loads TILE_COLs columns and TILE_ROW rows
// after reading a tile the row counter increase by TILE_ROWS
// the col counter reset after reading TILE_COL elements
const int base_row = ((blockIdx.x*TILE_COLS)/tiledCols)*TILE_ROWS;
// col increases by TILE_SIZE for each block and wraps back to 0 after tiledCols is reached
const int base_col = (blockIdx.x*TILE_COLS) % tiledCols;
const int base_idx = (base_row*cols) + base_col;
const int items_per_load = ITEMS_PER_THREAD*THREADS;
typedef cub::BlockLoad<T, THREADS, ITEMS_PER_THREAD, cub::BLOCK_LOAD_VECTORIZE> LoadT;
typedef cub::BlockReduce<float, THREADS> BlockRowReduce;
typedef cub::BlockReduce<int, THREADS> BlockRowSum;
typedef cub::BlockExchange<float, THREADS, ITEMS_PER_THREAD> BlockExchange;
__shared__ union {
typename BlockExchange::TempStorage exchange;
typename BlockRowReduce::TempStorage rowreduce;
typename BlockRowSum::TempStorage rowsum;
typename LoadT::TempStorage loadt;
} temp_storage;
__shared__ float smem_row_absmax_values[ITEMS_PER_THREAD*THREADS];
__shared__ int smem_row_nnz_values[TILE_ROWS];
half local_data[ITEMS_PER_THREAD];
float local_data_fp32[ITEMS_PER_THREAD];
float local_col_absmax_values[ITEMS_PER_THREAD];
int local_row_nnz_count = 0;
float row_absmax = -FLT_MAX;
// 0. reset stats to -FLT_MAX
for(int j = 0; j < ITEMS_PER_THREAD; j++)
{
//smem_col_absmax_values[threadIdx.x + (j*THREADS)] = -FLT_MAX;
smem_row_absmax_values[threadIdx.x + (j*THREADS)] = -FLT_MAX;
smem_row_nnz_values[threadIdx.x + (j*THREADS)] = 0;
}
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
local_col_absmax_values[j] = -FLT_MAX;
__syncthreads();
int valid_items = cols - base_col > items_per_load ? items_per_load : cols - base_col;
int i = base_idx;
// we load row after row from the base_position
// 1. load row-by-row ITEMS_PER_THREAD (TILE_SIZE==THREADS*ITEMS_PER_THREAD)
for(int row = 0; row < TILE_ROWS; row++)
{
if(base_row+row >= rows){ break; }
local_row_nnz_count = 0;
i = base_idx + ((row)*cols);
// each thread gets data from the same column
__syncthreads();
LoadT(temp_storage.loadt).Load(&(A[i]), local_data, valid_items, __float2half(0.0f));
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
local_data[j] = fabsf(local_data[j]);
if(SPARSE_DECOMP)
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
{
if((float)local_data[j] >= nnz_threshold)
{
local_row_nnz_count += 1;
local_data[j] = 0.0f;
}
}
// 2. compute col max (per thread); store in smem due to register pressure
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
// take the col max for this row
// we use shared memory because register pressure is too high if we do this locally
//smem_col_absmax_values[threadIdx.x + (j*THREADS)] = fmaxf(smem_col_absmax_values[threadIdx.x + (j*THREADS)], __half2float(local_data[j]));
local_col_absmax_values[j] = fmaxf(local_col_absmax_values[j], __half2float(local_data[j]));
// 3. compute row max (per block); store in smem to accumulate full global mem transation
// this is slow as it uses extra registers, but we need this to be compatible with Kepler and Maxwell (no fp16 units)
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
local_data_fp32[j] = local_data[j];
2022-07-22 22:21:37 +00:00
__syncthreads();
2022-07-22 21:41:05 +00:00
row_absmax = (float)BlockRowReduce(temp_storage.rowreduce).Reduce(local_data_fp32, cub::Max());
if(SPARSE_DECOMP)
{
__syncthreads();
local_row_nnz_count = BlockRowSum(temp_storage.rowsum).Sum(local_row_nnz_count);
}
// we store the data temporarily in shared memory so we
// can execute a full atomic block transaction into global memory later
// we use a striped arrangement [0, 8, 16, 24, ..] for t0 for faster stores
if(threadIdx.x == 0)
{
smem_row_absmax_values[(row % ITEMS_PER_THREAD) + ((row/ITEMS_PER_THREAD)*ITEMS_PER_THREAD)] = row_absmax;
// each blockIdx.x process 16 rows and 64*4=256 columns -> we sum nnz over 256 columns and have 16 values per block
smem_row_nnz_values[row] = local_row_nnz_count;
}
__syncthreads();
}
// 4. store data via atomicMax
// to store col data efficienctly we need to rewrite the smem blocked data [0, 1, 2, 3...] for t0
// into a striped arangement: [0, 8, 16, 24, ..] for t0
__syncthreads();
BlockExchange(temp_storage.exchange).BlockedToStriped(local_col_absmax_values);
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
if(base_col+threadIdx.x+(j*THREADS) < cols)
{
float val = colStats[base_col+(threadIdx.x+(j*THREADS))];
if(val < local_col_absmax_values[j])
atomicMax(&colStats[base_col+(threadIdx.x+(j*THREADS))], local_col_absmax_values[j]);
}
for(int j = 0; j < ITEMS_PER_THREAD; j++)
if(base_row+threadIdx.x+(j*THREADS) < rows)
{
float val = rowStats[base_row+(threadIdx.x+(j*THREADS))];
if(val < smem_row_absmax_values[threadIdx.x+(j*THREADS)])
atomicMax(&rowStats[base_row+(threadIdx.x+(j*THREADS))], smem_row_absmax_values[threadIdx.x+(j*THREADS)]);
}
if(SPARSE_DECOMP)
if(threadIdx.x < TILE_ROWS)
nnz_count_row[blockIdx.x*TILE_ROWS+threadIdx.x+1] = smem_row_nnz_values[threadIdx.x];
}
template __global__ void kgetColRowStats<half, 64, 4, 16, 64*4, 0>(half * __restrict__ A, float *rowStats, float *colStats, int * nnz_count_row, float nnz_threshold, int rows, int cols, int tiledRows, int tiledCols);
template __global__ void kgetColRowStats<half, 64, 4, 16, 64*4, 1>(half * __restrict__ A, float *rowStats, float *colStats, int * nnz_count_row, float nnz_threshold, int rows, int cols, int tiledRows, int tiledCols);
#define MM_DEQUANT_CONST 6.200012e-05f //1.0f/(127.0f*127.0f)
2022-08-16 17:56:17 +00:00
template <int ITEMS_PER_THREAD, int SUBTILE_ROWS, int THREADS>__global__ void kdequant_mm_int32_fp16(int *__restrict__ const A, float *__restrict__ const rowStats, float *__restrict__ const colStats, half *out, float* newRowStats, float* newcolStats, half *__restrict__ const bias, const int numRows, const int numCols, const int tileCols, const int n)
2022-07-22 21:41:05 +00:00
{
// Strategy: To dequantize we need to load col/row statistics. This can be very expensive
2022-10-27 11:11:29 +00:00
// since different row/col stats need to be loaded with each thread.
2022-07-22 21:41:05 +00:00
// (1, bad algorithm) Loading 32 items per thread would only occur 1 row load, but this increases register pressure
2022-10-27 11:11:29 +00:00
// and would lead to low global load utilization.
2022-07-22 21:41:05 +00:00
// (2, bad algorithm) If each thread loads some columns and multiple rows one needs to do lot of row loads
// for each thread and this is duplicated by a factor of 32/num-cols-per-thread.
// (3, good algorithm) Combining (1) and (2) we use sub-tiles of size 32xk in shared memory per threadblock.
// This allows for efficient row/col loading from shared memory within the tile.
// We can run for example 32x128 sub-tiles and warp-strided loads of 4 elements so that each thread has
// the same col statistic but needs to load 4 row stats from shared memory. To prevent bank conflicts
// we use a block-striped shared memory config [1, 31, 63, 95] so no bank conflicts happen during the
2022-10-27 11:11:29 +00:00
// shared memory loads.
2022-07-22 21:41:05 +00:00
// data is in 32 column-tile major with tile width 32 columns and numRows rows
// L1. Load sub-tile row/col statistics. Each thread only holds 1 col, load rows into shared memory.
// L2. Load data in warp-striped arangement (t0 holds colidx [0, 0, 0, 0], rowidx [0, 1, 2, 3])
// C1. Compute val(row_stat*col_stat)/(127*127) (load 1/(127*127 into register))
// C2. Compute normalization values and store col values in register
// S1. Store C1 into 16-bit output
// S2. Store col/row statistics of new buffer in shared memory
// We allow for sub-tiles to span multiple col32 tiles. This is okay
// since the items per thread only rely on a single column statistic.
const int n_out = numRows*numCols;
int num_row_tiles = (numRows/SUBTILE_ROWS) + (numRows % SUBTILE_ROWS == 0 ? 0 : 1);
// we have tiles of size numRows*32, thus col only increases every numRows
// num_row_tiles is the tiles after which the column increases by 32
// blockIdx.x is the index of the current tile
int col = ((threadIdx.x % 32) + ((blockIdx.x/num_row_tiles)*32));
// base_row increases by SUBTILE_ROWS every block. It wraps back to zero once num_row_tiles is reached
int base_row = (blockIdx.x*SUBTILE_ROWS) % (num_row_tiles*SUBTILE_ROWS);
// SUBTILE_ROWS is independent from ITEMS_PER_THREAD is independent from THREADS
// subtiles have 32*SUBTILE_ROWS elements <= THREADS*ITEMS_PER_THREAD
// Total subtiles should be n/(32*SUBTILE_ROWS) where each subtile has SUBTILE_ROW*32/4 threads.
// For example for a 1024x1024 matrix with 128 SUBTILE_ROWS and 4 ITEMS_PER_THREAD we have
// 1024*1024/(128*32) = 256 tiles
// 256 tiles are 256*128*32/4 = 256*1024 threads
// 1. Figure out how index relates to the start of the sub-tile
// 2. Each thread < SUBTILE_ROWS calculates row index
// 3. Load striped and store in shared memory
int local_values[ITEMS_PER_THREAD];
half local_output[ITEMS_PER_THREAD];
float local_rowStats[ITEMS_PER_THREAD];
__shared__ float smem_rowStats[SUBTILE_ROWS];
typedef cub::BlockLoad<int, THREADS, ITEMS_PER_THREAD, cub::BLOCK_LOAD_DIRECT> LoadInt32;
typedef cub::BlockExchange<int, THREADS, ITEMS_PER_THREAD> ExchangeInt32;
__shared__ typename LoadInt32::TempStorage loadint32;
__shared__ typename ExchangeInt32::TempStorage exchangeint32;
// L1. Load sub-tile row/col statistics. Each thread only holds 1 col, load rows into shared memory.
float colStat = col >= numCols ? 0.0f : colStats[col];
2022-08-16 18:12:09 +00:00
float local_biasValue = ((bias == NULL) || (col >= numCols)) ? 0.0f : __half2float(bias[col]);
2022-07-22 21:41:05 +00:00
// no block loads for rows for now -- keep it simple
for(int j = threadIdx.x; j < SUBTILE_ROWS; j+=blockDim.x)
{
// todo: is this global mem access slow due to overlaps or does the L1 cache work well here?
int row = (base_row+j) % numRows; // wrap around
// each warp accesses the same element, for four consequitive elements
// todo: update description about striped shared memory, it is not needed
// rowidx: [0, 1, 2, 3...] and each warp reads ITEMS_PER_THREAD consequitive elements
smem_rowStats[j] = rowStats[row];
}
__syncthreads();
// each block processes SUBTILE_ROWS*32 elements
const int items_per_load = THREADS*ITEMS_PER_THREAD;
const int rows_per_load = items_per_load/32;
int subtile_base_row = (threadIdx.x / 32)*ITEMS_PER_THREAD; // row within the tile
int row_offset = 0;
// subtile_idx starts at the base_row*32 + the total offset for a full numRow*32 tile is passed
int subtile_start = (blockIdx.x/num_row_tiles)*(numRows*32) + (base_row*32);
for(int subtile_idx = subtile_start; subtile_idx < subtile_start + (SUBTILE_ROWS*32); subtile_idx+=items_per_load)
{
int valid_rows = numRows - (base_row+row_offset) > rows_per_load ? rows_per_load : numRows - (base_row+row_offset);
int valid_items = valid_rows*32;
if(valid_items <= 0) // the sub-tile might have more elements than the tile itself
break;
// L2. Load data in warp-striped arangement (t0 holds colidx [0, 0, 0, 0], rowidx [0, 1, 2, 3])
LoadInt32(loadint32).Load(&(A[subtile_idx]), local_values, valid_items, 0);
ExchangeInt32(exchangeint32).BlockedToWarpStriped(local_values, local_values);
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
local_rowStats[j] = smem_rowStats[subtile_base_row+row_offset+j];
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
2022-08-16 18:12:09 +00:00
local_output[j] = __float2half((local_values[j]*MM_DEQUANT_CONST*local_rowStats[j]*colStat) + local_biasValue);
2022-07-22 21:41:05 +00:00
//absmax_col = fmax(fabsf(local_output[j]), absmax_col);
// we store data in row major
// to store data efficiently, we want to use block exchange: [0, 32, 64, 92] -> [0, 1, 2, 3]
// so that each thread holds ITEMS_PER_THREAD consecutive items for each row
// this way throughput into storage is increased by a factor of ~2x
// for now we use a simple store
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
{
int outIdx = col + ((base_row+subtile_base_row+row_offset+j)*numCols);
if(outIdx< n_out && col < numCols)
out[outIdx] = local_output[j];
}
row_offset += rows_per_load;
}
}
template <int THREADS, int ITEMS_PER_THREAD, int TILE_ROWS, int TILE_COLS, int SPARSE_DECOMP> __global__ void kDoubleRowColQuant(half *__restrict__ const A, float *__restrict__ const rowStats, float * __restrict__ const colStats, char *out_col_normed, char *out_row_normed, int *rowidx, int *colidx, half *val, int * __restrict__ nnz_block_ptr, float threshold, int rows, int cols, int tiledCols)
{
// assumes TILE_SIZE == THREADS*ITEMS_PER_THREAD
// Each thread reads the same column but multiple rows
// Rows are loaded in shared memory and access is shared across the threadblock (broadcast)
// 0. Load row stats data into shared memory; load col stat (1 fixed per thread)
// 1. Load data row by row (should be at least with TILE_SIZE = 512)
// 2. quantize data with row/col stats
// 3. Store data (TILE_SIZE = 512 is a bit slow, but should still be close enough to good performance)
// each block loads TILE_COLs columns and TILE_ROW rows
// after reading a tile the row counter increase by TILE_ROWS
// the col counter reset after reading TILE_COL elements
const int base_row = ((blockIdx.x*TILE_COLS)/tiledCols)*TILE_ROWS;
// col increases by TILE_SIZE for each block and wraps back to 0 after tiledCols is reached
const int base_col = (blockIdx.x*TILE_COLS) % tiledCols;
const int base_idx = (base_row*cols) + base_col;
const int items_per_load = ITEMS_PER_THREAD*THREADS;
typedef cub::BlockLoad<half, THREADS, ITEMS_PER_THREAD, cub::BLOCK_LOAD_VECTORIZE> LoadHalf;
__shared__ typename LoadHalf::TempStorage loadhalf;
typedef cub::BlockStore<char, THREADS, ITEMS_PER_THREAD, cub::BLOCK_STORE_VECTORIZE> StoreInt8;
__shared__ typename StoreInt8::TempStorage storeint8;
__shared__ float smem_row_stats[TILE_ROWS];
__shared__ unsigned int smem_nnz_row_idx[TILE_ROWS];
half local_data[ITEMS_PER_THREAD];
float local_col_stats[ITEMS_PER_THREAD];
char local_quantized_data[ITEMS_PER_THREAD];
// 0. Load row stats data into shared memory; load col stat (1 fixed per thread)
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
if(base_col+(threadIdx.x*ITEMS_PER_THREAD) + j < cols)
local_col_stats[j] = __fdividef(127.0f, colStats[base_col+(threadIdx.x*ITEMS_PER_THREAD)+j]);
for(int i = threadIdx.x; i < TILE_ROWS; i+=blockDim.x)
{
if(base_row + i < rows)
smem_row_stats[i] = rowStats[base_row+i];
if(SPARSE_DECOMP)
smem_nnz_row_idx[i] = nnz_block_ptr[(TILE_ROWS*blockIdx.x) + i];
}
__syncthreads();
// we load row after row from the base_position
// 1. Load data row by row (should be at least with TILE_SIZE = 512)
for(int row = 0; row < TILE_ROWS; row++)
{
if(base_row + row >= rows){ break; }
int i = base_idx + (row*cols);
int valid_items = cols - base_col > items_per_load ? items_per_load : cols - base_col;
LoadHalf(loadhalf).Load(&(A[i]), local_data, valid_items, 0.0f);
float row_stat = __fdividef(127.0f, smem_row_stats[row]);
// 2. quantize data with row/col stats
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
{
// we already pre-normalized the col/row stat:
// what this does is float/absmax*127 = int8
if(SPARSE_DECOMP)
{
if(fabsf((float)local_data[j]) >= threshold)
{
local_quantized_data[j] = 0;
int old_idx = atomicInc(&smem_nnz_row_idx[row], UINT_MAX);
rowidx[old_idx] = base_row+row;
colidx[old_idx] = base_col+(threadIdx.x*ITEMS_PER_THREAD)+j;
val[old_idx] = local_data[j];
}
else
{
local_quantized_data[j] = (char)(rintf(__half2float(local_data[j])*row_stat));
}
}
else
local_quantized_data[j] = (char)(rintf(__half2float(local_data[j])*row_stat));
}
StoreInt8(storeint8).Store(&(out_row_normed[i]), local_quantized_data, valid_items);
// 2. quantize data with row/col stats
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
{
// we already pre-normalized the col/row stat:
// what this does is float/absmax*127 = int8
local_quantized_data[j] = (char)(rintf(__half2float(local_data[j])*local_col_stats[j]));
}
__syncthreads();
StoreInt8(storeint8).Store(&(out_col_normed[i]), local_quantized_data, valid_items);
}
}
template <int THREADS, int ITEMS_PER_THREAD, int TILE_ROWS, int TILE_COLS, int TRANSPOSE, int FORMAT> __global__ void kTransformRowToFormat(char *__restrict__ const A, char *out, int rows, int cols, int tiledCols, int outRows, int outCols)
{
// 0. Load data into 32*32 shared memory tiles
// 1. transpose / reorder in shared memory
// 2. store
// COL32 FORMAT:
// rows*32 tiles
// TURING FORMAT:
// 8*32 tiles with 4*4 subtiles
// the 8*32 subtile has first all 4*4 subtiles of even rows (max 4*4*4 = 64 elements)
// the subsequent 4*4 subtiles are for all odd rows if some rows columns are empty the values are zero
// the tile repeats again after the 8*32 tile in a major column order, meaning: (next 8 rows are A[8:16, 0:32])
// the next tile is the next 8 rows for the same 32 columns. Once all rows are finished, the column
// index increases by 32
// AMPERE FORMAT:
// 32*32 tiles with 8*32 subtiles. The rows are interleaved in pairs of two rows with offset of 8 between pairs of two rows:
// row idx (each number stands for 32 values): [0 1 8 9 16 17 24 25] [2 3 10 11 18 19 26 27]...
// the tiles are column-major ordered, so after 1024*1024 values we process: A[32:64, 0:32]
// To have efficient loads and stores if we transpose we need 128 consequitive bytes which at 1 byte are 128 values
2022-10-27 11:11:29 +00:00
// As such we need:
2022-07-22 21:41:05 +00:00
// at least 32*4 shared memory tiles for col32; preferably 32*32
// at least 32*6 shared memory tiles for col32_ampere: preferably 32*32
// at least 32*8 shared memory tiles for col4_turing: preferably 32*32
// for efficient loading of row major we need to load 128 elements and repeat this 32 items
// this would imply a 32x128 shared memory tile -> 4kb
// It is more efficient to have more than 1 warp, so with 64 threads we need 32x128 -> 8 kb
// we have 64k sharded mem per SM in Turing which is 8 blocks per SM which is 2*8 = 32 warps = 100% occupancy
// for turing and 50% for A100 and 75% for RTX 30s / A40 which is probably good enough
// register pressure should be low with: 8 registers from local memoryh per block and 64 registers per SM
2022-10-27 11:11:29 +00:00
//
2022-07-22 21:41:05 +00:00
// to make the shared memory work with that occupancy we might need to union the block loads/stores
// each block loads TILE_COLs columns and TILE_ROW rows
// after reading a tile the row counter increase by TILE_ROWS
// the col counter reset after reading TILE_COL elements
const int base_row = ((blockIdx.x*TILE_COLS)/tiledCols)*TILE_ROWS;
// col increases by TILE_SIZE for each block and wraps back to 0 after tiledCols is reached
const int base_col = (blockIdx.x*TILE_COLS) % tiledCols;
const int base_idx = (base_row*cols) + base_col;
// we load 128 bytes per warp with
// 32 rows for transposes that fill col32 types
// so that we can have contiguous stores
__shared__ char smem_data[32*33*ITEMS_PER_THREAD];
char local_data[ITEMS_PER_THREAD];
typedef cub::BlockExchange<char, THREADS, ITEMS_PER_THREAD> BlockExchange;
// we load row after row from the base_position
// Load data row by row
int warps = blockDim.x/32;
int warp_id = threadIdx.x/32;
int warp_lane = threadIdx.x % 32;
int offset = 0;
int smem_row = 0;
// each warp loads one row of 128 bytes
for(int row = warp_id; row < TILE_ROWS; row+=warps)
{
int i = base_idx + (row*cols);
// we load up to 128 bytes/items per load
int valid_items = cols - base_col > 32*ITEMS_PER_THREAD ? 32*ITEMS_PER_THREAD : cols - base_col;
// 0. Load data into 32*32 shared memory tiles
if(base_row + row < rows)
{
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
{
int col_idx = warp_lane+(j*32);
if(col_idx < valid_items)
local_data[j] = A[i+col_idx];
else
local_data[j] = 0;
}
}
else
{
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
local_data[j] = 0;
}
if(TRANSPOSE)
{
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
{
int local_col = (32*j)+warp_lane;
//int local_row = row;
// store as 256x32
smem_data[(local_col*33) + row] = local_data[j];
}
}
else
{
// treat smem as 32x256, that is 32 rows and 256 columns
#pragma unroll ITEMS_PER_THREAD
for(int j = 0; j < ITEMS_PER_THREAD; j++)
smem_data[row*32*ITEMS_PER_THREAD + (warp_lane) + (j*32)] = local_data[j];
}
smem_row += warps;
// 1. transpose / reorder in shared memory
if(smem_row % 32 == 0)
{
smem_row = 0;
__syncthreads();
for(int subrow = warp_id; subrow < 32; subrow+=warps)
{
for(int j = 0; j < ITEMS_PER_THREAD; j++)
{
switch(FORMAT)
{
2022-10-27 11:11:29 +00:00
case COL32:
2022-07-22 21:41:05 +00:00
if(TRANSPOSE)
{
// data lies in shared memory in the following way:
// row0 [col0 col1 ... col31]
// row1 [col0 col1 ... col31]
// ...
//
// As such we read consequtive entries with 256 threads (8rows x 32 columns)
// as j increase, the row increase by a factor of 8
// We load 8 rows per subrow loop, and subrow increase by 8 per loop
// so we have an offset of 8 rows every loop or (subrow/warps)*8 = (subrow/8)*8
const int jrow = j*ITEMS_PER_THREAD; // 8 rows per j
const int subrow_loop_row = (subrow/warps)*ITEMS_PER_THREAD*ITEMS_PER_THREAD; // 8 rows per j; 8j per subrow loop (subrow/warps)
//const int local_row = warp_id; // each warp_id is one row
//const int block_row = base_col; // block offset for row
//const int local_col = warp_lane
//const int global_col = base_row; // block offset for col
if((base_col + subrow_loop_row + jrow + warp_id < outRows) && (base_row+warp_lane < rows))
{
// each row hae 32 columns and is offset by 1 to prevent bank conflict during storage into smem
char data = smem_data[(subrow_loop_row + jrow + warp_id)*33 + warp_lane];
// each 32 columns we have new tile
// each tile has size outRows*32 and base_row is done in increments of 32
2022-10-27 11:11:29 +00:00
offset = base_row*outRows;
2022-07-22 21:41:05 +00:00
out[offset + (base_col + jrow + subrow_loop_row)*32 + threadIdx.x] = data;
}
}
else
{
if(((base_row+subrow) < rows) && (base_col+(j*32)+warp_lane < outCols))
{
offset = (base_col/32)*(32*rows);
char data = smem_data[(subrow*32*ITEMS_PER_THREAD) + (j*32) + warp_lane];
out[offset+(base_row+subrow)*32 + ((j)*rows*32)+warp_lane] = data;
}
}
break;
case COL_TURING:
// TURING FORMAT:
// 8*32 tiles with 4*4 subtiles
// the 8*32 subtile has first all 4*4 subtiles of even rows (max 4*4*4 = 64 elements)
// the subsequent 4*4 subtiles are for all odd rows if some rows columns are empty the values are zero
// the tile repeats again after the 8*32 tile in a major column order, meaning: (next 8 rows are A[8:16, 0:32])
// the next tile is the next 8 rows for the same 32 columns. Once all rows are finished, the column
// index increases by 32
//
// [0 0 0 0, 2 2 2 2, 4 4 4 4, 6 6 6 6, 0 0 0 0 ...]
if(TRANSPOSE)
{
const int jrow = j*ITEMS_PER_THREAD; // 8 rows per j
const int subrow_loop_row = (subrow/warps)*ITEMS_PER_THREAD*ITEMS_PER_THREAD; // 8 rows per j; 8j per subrow loop (subrow/warps)
//const int local_row = warp_id; // each warp_id is one row
//const int block_row = base_col; // block offset for row
//const int local_col = warp_lane
//const int global_col = base_row; // block offset for col
if((base_col + subrow_loop_row + jrow + warp_id < outRows) && (base_row+warp_lane < rows))
{
// each row hae 32 columns and is offset by 1 to prevent bank conflict during storage into smem
char data = smem_data[(subrow_loop_row + jrow + warp_id)*33 + warp_lane];
// each 32 columns we have new tile
// each tile has size 8*32 = 256 elements offset
// for each row offset of 8 we increaes the tile first
// after all rows are exhausted, we increase the col
int row_offset = ((base_col+jrow+subrow_loop_row+warp_id)/8)*256; // global_row+jrow+subrow_loop_row+local_row, increase tile(=256) every 8 rows
// we increase by row_tile_column every 32 columns
// base_row increase in increments of 32
//int row_tile_column = 256*outRows/8; // there are outRows/8 row tiles, and each tile is 256 elements
2022-10-27 11:11:29 +00:00
//int col_offset = (base_row/32)*row_tile_column;
2022-07-22 21:41:05 +00:00
// -> we can remove the divisions to speed up compute since outRows is always a multiple of 8
// 256*outRows/8*base_row/32 = outRows*base_row
int col_offset = outRows*base_row;
offset = row_offset+col_offset;
// since we process even number of rows with each j (8) and with each subrow (8j) we can determine
// odd or even rows with the warp_id (each warp processes one row)
// the col is warp_lane (max 32 columns per row) and the row warp_id
if(warp_id % 2 == 1)
// odd
offset += 128 + (warp_lane/4)*16 + (warp_lane%4) + (((warp_id%8)-1)*2);
else
// even
offset += 0 + (warp_lane/4)*16 + (warp_lane%4) + ((warp_id%8)*2);
out[offset] = data;
}
}
else
{
if(((base_row+subrow) < rows) && (base_col+(j*32)+warp_lane < outCols))
{
char data = smem_data[(subrow*32*ITEMS_PER_THREAD) + (j*32) + warp_lane];
// set offset designates the tile offset among the 8*32 tiles
// we first increase rows and then columns. Since we load 128 columns at once
// we increase the offset by outRows*32 every 32 columns
// additionally, we increase the offset by 8*32=256 every 8 rows
offset = ((base_col+(j*32))/32)*outRows*32 + (((base_row+subrow)/8)*256); // global offset (8x32 tile)
// first 4 rows are reserved for even rows, [0, 2, 4, 6], the next 4 for odd
// each of these has 32 values in total for 32*4 = 128 as offset if odd
// every set of 4 columns increases the total offset by 16
// each even row increase the offset by 4, for example row 2 is offset by 4, 4 by 6 etc so: subrow/2*4 = subrow*2
// this happends every 8 rows anew (subrow % 8)
// one writes 4 columns at once that is (col % 4) for the particular index in the subtile
int subcol = warp_lane;
2022-10-27 11:11:29 +00:00
2022-07-22 21:41:05 +00:00
// add local offset (4x4 sub-tile)
if(subrow % 2 == 1)
// odd
offset += 128 + (subcol/4)*16 + (subcol%4) + (((subrow%8)-1)*2);
else
// even
offset += 0 + (subcol/4)*16 + (subcol%4) + ((subrow%8)*2);
out[offset] = data;
}
}
break;
case COL_AMPERE:
// AMPERE FORMAT:
// 32*32 tiles with 8*32 subtiles. The rows are interleaved in pairs of two rows with offset of 8 between pairs of two rows:
// row idx (each number stands for 32 values): [0 1 8 9 16 17 24 25] [2 3 10 11 18 19 26 27]...
// the tiles are column-major ordered, so after 1024*1024 values we process: A[32:64, 0:32]
if(TRANSPOSE)
{
const int jrow = j*ITEMS_PER_THREAD; // 8 rows per j
const int subrow_loop_row = (subrow/warps)*ITEMS_PER_THREAD*ITEMS_PER_THREAD; // 8 rows per j; 8j per subrow loop (subrow/warps)
//const int local_row = warp_id; // each warp_id is one row
//const int block_row = base_col; // block offset for row
//const int local_col = warp_lane
//const int global_col = base_row; // block offset for col
if((base_col + subrow_loop_row + jrow + warp_id < outRows) && (base_row+warp_lane < rows))
{
// each row hae 32 columns and is offset by 1 to prevent bank conflict during storage into smem
char data = smem_data[(subrow_loop_row + jrow + warp_id)*33 + warp_lane];
// each 32 columns we have new tile
// each tile has size 32*32 = 1024 elements offset
// for each row offset of 32 we increaes the tile first
// after all rows are exhausted, we increase the col
int row_offset = ((base_col+jrow+subrow_loop_row+warp_id)/32)*1024; // global_row+jrow+subrow_loop_row+local_row, increase tile(=256) every 8 rows
// we increase by row_tile_column every 32 columns
// base_row increase in increments of 32
//int row_tile_column = 1024*outRows/32; // there are outRows/32 row tiles, and each tile is 1024 elements
2022-10-27 11:11:29 +00:00
//int col_offset = (base_row/32)*row_tile_column;
2022-07-22 21:41:05 +00:00
// -> we can remove the divisions to speed up compute since outRows is always a multiple of 8
// 1024*outRows/32*base_row/32 = outRows*base_row
int col_offset = outRows*base_row;
offset = row_offset+col_offset;
// same as in the non-transpose case (see below)
// the difference is that now rows = cols
// in this case warp_id = subrow
// [0 1 8 9 16 17 24 25] [2 3 10 11 18 19 26 27]...
// subrow % 8 -> [0,1] in tile0, [2, 3] in tile 1 etc
// subrow % 2 -> 0 for 1st row in the pair, 1 for the 2nd row
// every 2 rows, the offset increases by two [0, 1, 8, 9...]
// every 2 rows, the row index increase by 8 [0, 1, 8, 9...]
int local_row = (jrow + warp_id) % 32; // offset for row > 32 is already calculated into row_offset
int ampere_row = ((local_row % 8)/2)*8 + (local_row/8)*2 + (local_row % 2);
// global offset + row with 32 cols each + 32 cols per j + col_idx=warp_lane
out[offset + (ampere_row*32) + warp_lane] = data;
}
}
else
{
if(((base_row+subrow) < rows) && (base_col+(j*32)+warp_lane < outCols))
{
char data = smem_data[(subrow*32*ITEMS_PER_THREAD) + (j*32) + warp_lane];
// set offset designates the tile offset among the 32*32 tiles
// we first increase rows and then columns. Since we load 128 columns at once
// we increase the offset by outRows*32 every 32 columns
// additionally, we increase the offset by 32*32=1024 every 32 rows
offset = ((base_col+(j*32))/32)*outRows*32 + (((base_row+subrow)/32)*1024); // global offset (32x32 tile)
// [0 1 8 9 16 17 24 25] [2 3 10 11 18 19 26 27]...
// subrow % 8 -> [0,1] in tile0, [2, 3] in tile 1 etc
// subrow % 2 -> 0 for 1st row in the pair, 1 for the 2nd row
// every 2 rows, the offset increases by two [0, 1, 8, 9...]
// every 2 rows, the row index increase by 8 [0, 1, 8, 9...]
int local_row = ((subrow % 8)/2)*8 + (subrow/8)*2 + (subrow % 2);
// global offset + row with 32 cols each + 32 cols per j + col_idx
out[offset + (local_row*32) + warp_lane] = data;
}
}
break;
}
}
}
}
}
}
2023-04-26 00:15:51 +00:00
#define DENORM 1.0f/127.0f
2022-07-22 21:41:05 +00:00
#define MAX_SPARSE_COUNT 32
#define SMEM_SIZE 8*256
2022-10-27 11:11:29 +00:00
template <typename T, int SPMM_ITEMS, int BITS>
2022-07-26 02:30:37 +00:00
__global__ void kspmm_coo_very_sparse_naive(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, T *B, half *out, float * __restrict__ const dequant_stats, int nnz, int rowsA, int rowsB, int colsB)
2022-07-22 21:41:05 +00:00
{
// 0. load balancing: We process rows with most columns first (count_vec)and we process one row per block
// If a block finishes, the next one is scheduled. Since the last blocks like have fewer
// elements they finish faster "fillin up" the gaps left by larger blocks
// without tensor cores
// 1. use rowidx_length to find what to load (as many blocks as there are rows)
// 2. Load A into registers
// 3. each warp loads all required rows of B but each warp is offset by k
// 4. Do mma operations that accumulate into registers
// 5. Each warp stores its output row into matrix C
const int count = max_count[blockIdx.x];
const int local_max_idx = max_idx[blockIdx.x];
const int offset = local_max_idx == 0 ? 0 : offset_rowidx[local_max_idx-1];
const int local_row_idx = rowidx[offset];
const int warp_id = threadIdx.x / 32;
const int warp_idx = threadIdx.x % 32;
const int warp_offset = (warp_id*32)*SPMM_ITEMS;
const int num_items = BITS == 8 ? 8 : 8;
int idx_col_B = warp_offset;
int local_idx_col_B_offset = 0;
half local_valA[MAX_SPARSE_COUNT];
int local_colidxA[MAX_SPARSE_COUNT];
half local_valC[SPMM_ITEMS];
T local_valsB[num_items];
half local_valOut[num_items];
// 128 byte loads per warp == 4 bytes per thread
// 2. Load A into registers
for(int j = 0; j < MAX_SPARSE_COUNT; j++)
{
local_valA[j] = j < count ? values[offset+j] : __float2half(0.0f);
local_colidxA[j] = j < count ? colidx[offset+j] : 0;
}
// each thread processes SPMM_ITEMS=32 per iteration. We have 256 threads. 32*256=x192
// we expect each warp to be SPMM_ITEMS*32 apart
// we have a total of 128 bytes for the bank with a bank size of 4 bytes
// added 3 bytes = 6 values between warps should reduce bank conflicts
__shared__ half smem_dequant_stats[SMEM_SIZE];
while(idx_col_B < colsB)
{
if(dequant_stats != NULL)
{
for(int i = threadIdx.x; i < SMEM_SIZE; i+=blockDim.x)
if((idx_col_B+i-local_idx_col_B_offset) < colsB)
2022-07-26 02:30:37 +00:00
smem_dequant_stats[i] = dequant_stats[idx_col_B+i-local_idx_col_B_offset];
2022-07-22 21:41:05 +00:00
__syncthreads();
}
#pragma unroll SPMM_ITEMS
for(int j = 0; j < SPMM_ITEMS; j++)
local_valC[j] = 0.0f;
#pragma unroll
for(int i = 0; i < count; i++)
{
// 3. each warp loads all required rows of B but each warp is offset by k
int row_offset = colsB*local_colidxA[i];
#pragma unroll SPMM_ITEMS
for(int j = 0; j < SPMM_ITEMS; j+=num_items)
{
// 4. Multiply the tile -> accumulate outputs in shared memory until 128 bytes it reached
int idx = idx_col_B + (warp_idx*SPMM_ITEMS) + j;
if(idx >= colsB){ break; }
if((idx+num_items < colsB))
{
if(BITS == 8)
reinterpret_cast<float2(&)[num_items]>(local_valsB)[0] = reinterpret_cast<float2*>(B)[(row_offset+ idx)/num_items];
else
reinterpret_cast<float4(&)[num_items]>(local_valsB)[0] = reinterpret_cast<float4*>(B)[(row_offset+ idx)/num_items];
}
else
{
#pragma unroll num_items
for(int k = 0; k < num_items; k++)
if(idx+k < colsB)
local_valsB[k] = B[row_offset+idx+k];
else
local_valsB[k] = 0.0f;
}
#pragma unroll num_items
for(int k = 0; k < num_items; k++)
{
if(BITS == 8 && dequant_stats != NULL)
// we do texture cache reads (__ldg) on dequant_stats which should be super fast
{
float valB = local_valsB[k];
float valA = local_valA[i];
if(valB != 0.0 && valA != 0.0)
2023-04-26 00:15:51 +00:00
local_valC[j+k] = (float)local_valC[j+k] + ((float)smem_dequant_stats[idx+k-local_idx_col_B_offset])*DENORM*valB*valA;
2022-07-22 21:41:05 +00:00
}
else
local_valC[j+k] = (float)local_valC[j+k] + (float)local_valsB[k]*(float)local_valA[i];
}
}
}
int idx_row_C = (colsB*local_row_idx);
#pragma unroll SPMM_ITEMS
for(int j = 0; j < SPMM_ITEMS; j+=num_items)
{
//int idx_col_C = idx_col_B + (32*j) + warp_idx;
int idx_col_C = idx_col_B + warp_idx*SPMM_ITEMS + j;
int idx_val = idx_col_C + idx_row_C;
if(idx_col_C +num_items < colsB)
{
// load outputs to do inplace addition
reinterpret_cast<float4(&)[num_items/4]>(local_valOut)[0] = reinterpret_cast<float4*>(out)[idx_val/num_items];
#pragma unroll num_items
for(int k = 0; k < num_items; k++)
local_valC[(j/num_items) + k] = (float)local_valC[(j/num_items) + k] + (float)local_valOut[k];
2022-10-27 11:11:29 +00:00
2022-07-22 21:41:05 +00:00
reinterpret_cast<float4*>(out)[idx_val/num_items] = reinterpret_cast<float4(&)[num_items]>(local_valC)[j/num_items];
}
else
{
#pragma unroll num_items
for(int k = 0; k < num_items; k++)
if(idx_col_C + k < colsB)
out[idx_val+k] = (float)out[idx_val+k]+(float)local_valC[j+k];
}
}
idx_col_B += blockDim.x*SPMM_ITEMS;
local_idx_col_B_offset += blockDim.x*SPMM_ITEMS;
2022-10-27 11:11:29 +00:00
}
2022-07-22 21:41:05 +00:00
}
2022-07-27 00:39:30 +00:00
template <int FORMAT> __global__ void kExtractOutliers(char *A, int *idx, char *out, int idx_size, int rowsA, int colsA, int tiledRowsA, int tiledColsA)
2022-10-27 11:11:29 +00:00
{
2022-07-27 00:39:30 +00:00
int local_colidx = idx[blockIdx.x];
if(FORMAT==COL_TURING)
{
// TURING FORMAT:
// 8*32 tiles with 4*4 subtiles
// the 8*32 subtile has first all 4*4 subtiles of even rows (max 4*4*8 = 128 elements)
// the subsequent 4*4 subtiles are for all odd rows if some rows columns are empty the values are zero
// the tile repeats again after the 8*32 tile in a major column order, meaning: (next 8 rows are A[8:16, 0:32])
// the next tile is the next 8 rows for the same 32 columns. Once all rows are finished, the column
// index increases by 32
// columns are grouped in increments of 4, meaning that one has the following rows and columns
// rows: [0 0 0 0, 2 2 2 2, 4 4 4 4, 6 6 6 6, 0 0 0 0 ...]
// cols: [0 1 2 3, 0 1 2 4, 0 1 2 3, 0 1 2 3, 4 5 6 7 ...]
// each thread reads 1 element = 1 row
for(int row = threadIdx.x; row < rowsA; row+= blockDim.x)
{
int offset_per_col_tile = ((rowsA+7)/8)*32*8;
int tile_offset_rows = (row/8)*32*8;
int tile_offset_cols = (local_colidx/32)*offset_per_col_tile;
int offset = 0;
int subtile_col_idx = local_colidx%32;
int subtile_row_idx = row % 8;
if(row % 2 == 1)
offset += 128 + (subtile_col_idx/4)*16 + (subtile_col_idx%4) + ((subtile_row_idx-1)*2);
else
// even
offset += 0 + (subtile_col_idx/4)*16 + (subtile_col_idx%4) + (subtile_row_idx*2);
offset += tile_offset_rows + tile_offset_cols;
2022-07-27 01:15:51 +00:00
char val = A[offset];
2022-07-27 00:39:30 +00:00
int out_idx = (row*idx_size) + blockIdx.x;
2022-07-27 01:15:51 +00:00
out[out_idx] = val;
2022-07-27 00:39:30 +00:00
}
2022-07-27 01:15:51 +00:00
}
else if(FORMAT == COL_AMPERE)
{
2022-07-27 00:39:30 +00:00
2022-07-27 01:15:51 +00:00
for(int row = threadIdx.x; row < rowsA; row+= blockDim.x)
{
// we got 32x32 tiles and we use the magic equation from the cublasLt doc to get the element
// within each tile.
int offset_per_col_tile = ((rowsA+31)/32)*32*32;
int tile_offset_rows = (row/32)*32*32;
int tile_offset_cols = (local_colidx/32)*offset_per_col_tile;
int subtile_col_idx = local_colidx%32;
int subtile_row_idx = row % 32;
// this magic is taken from the cublasLt doc (search for COL32)
int offset = (((subtile_row_idx%8)/2*4+subtile_row_idx/8)*2+subtile_row_idx%2)*32+subtile_col_idx;
offset += tile_offset_cols + tile_offset_rows;
char val = A[offset];
int out_idx = (row*idx_size) + blockIdx.x;
out[out_idx] = val;
}
2022-07-27 00:39:30 +00:00
}
2022-10-27 11:11:29 +00:00
}
2022-07-26 19:12:38 +00:00
2023-04-25 23:13:43 +00:00
2023-04-27 00:12:34 +00:00
//template <int QUANT_TYPE, typename INPT, typename COMPT, typename OUTT> __global__ void kMatmul_inference_4bit(INPT *A, unsigned char *B, OUTT *out, int lda, int ldb, int rowsA, int colsA, int colsB)
//{
//// element-wise kernel
//// 1. Load batch x k into registers
//// 2. Load k x k into registers
//// 3. dequantize and store in second pair of k x k
//// 4. matmul
//// 5. sum with cub
//// 6. store outputs
//// TC kernel
//// use k warps per thread block
//// 1. threadblock use read-only cache to read in register tile for A into shared memory
//// 2. each warp loops over shared memory tiles of A of size 8x16 and loads them into fragments
//// 3. each warp reads a segment of values 16x32 from B
//// 4. do dequantization from register of B into second pair of registers
//// 5. store (4) into fragment
//// 6. matmul aggregate into fragment C
//// 7. aggreecate files of C into shared memroy block C
//// 8. sum (7)
//// 9. write outputs to matmul output matrix
//}
2023-04-25 23:13:43 +00:00
2023-04-30 04:52:47 +00:00
template <typename T, typename TCAST, int ITEMS> __device__ inline void vector_load(T *local, T * __restrict__ const buffer, int idx, int limit_base, int limit, float zero_value = 0.0f)
2023-04-29 05:10:42 +00:00
{
if(limit_base + ITEMS <= limit)
reinterpret_cast<TCAST*>(local)[0] = reinterpret_cast<TCAST*>(buffer)[idx/ITEMS];
else
{
for(int k = 0; k < ITEMS; k++)
{
if(limit_base + k < limit)
local[k] = buffer[idx+k];
else
2023-04-30 04:52:47 +00:00
local[k] = (T)zero_value;
2023-04-29 05:10:42 +00:00
}
}
}
2023-07-04 01:45:38 +00:00
#define WARPS 3
2023-04-29 05:10:42 +00:00
template <typename T, int BITS, int THREADS> __global__ void gemm_device(int M, int N, int K, T * __restrict__ const A, T* B, T * out, int lda, int ldb, int ldc)
2023-04-25 23:15:44 +00:00
{
2023-05-31 02:06:59 +00:00
#if __CUDA_ARCH__ >= 750
using namespace nvcuda;
2023-05-01 23:38:09 +00:00
int col_offset = blockIdx.x *32;
2023-05-01 02:12:42 +00:00
const int warp_id = threadIdx.x / 32;
2023-05-01 01:19:30 +00:00
const int half_warp_id = threadIdx.x / 16;
const int half_warp_lane = threadIdx.x % 16;
2023-05-01 15:21:12 +00:00
const int batch_size_warps = (WARPS-1)*2;
2023-05-02 23:15:38 +00:00
const int val_per_iter = blockDim.x-32;
2023-04-25 23:15:44 +00:00
2023-05-02 23:15:38 +00:00
T local_A[4];
T local_B[128];
2023-04-29 00:19:02 +00:00
2023-05-02 18:38:11 +00:00
const int a_tile_offset = 16;
2023-05-02 15:58:59 +00:00
const int b_tile_offset = (16*32 + 16);
2023-05-01 01:19:30 +00:00
2023-05-02 18:38:11 +00:00
__shared__ T smem_A[8*16 + (2*16*(batch_size_warps-1))];
2023-05-01 23:38:09 +00:00
__shared__ T smem_B[2*batch_size_warps*16*32 + (2*16*(batch_size_warps-1))];
2023-05-02 18:38:11 +00:00
//__shared__ T smem_C[8*32];
2023-04-29 00:19:02 +00:00
2023-05-01 23:38:09 +00:00
wmma::fragment<wmma::matrix_a, 8, 32, 16, half, wmma::row_major> a_frag;
wmma::fragment<wmma::matrix_b, 8, 32, 16, half, wmma::col_major> b_frag;
wmma::fragment<wmma::accumulator, 8, 32, 16, half> c_frag;
2023-05-01 00:43:02 +00:00
wmma::fill_fragment(c_frag, 0.0f);
2023-05-01 15:21:12 +00:00
int ticktock = 0;
int idx = 0 + threadIdx.x;
2023-05-02 19:10:32 +00:00
int loaded_values = 0;
2023-05-01 15:21:12 +00:00
// prefetch
if(idx < K && warp_id < (WARPS-1))
2023-04-29 04:29:40 +00:00
{
2023-05-02 19:10:32 +00:00
if(loaded_values == 0)
{
local_A[0] = A[idx];
2023-05-02 23:15:38 +00:00
local_A[1] = A[idx+(1*val_per_iter)];
local_A[2] = A[idx+(2*val_per_iter)];
local_A[3] = A[idx+(3*val_per_iter)];
2023-04-29 00:19:02 +00:00
2023-05-02 19:10:32 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
{
local_B[col] = B[(col_offset+col)*ldb+idx];
2023-05-02 23:15:38 +00:00
local_B[col+32] = B[(col_offset+col)*ldb+idx+(1*val_per_iter)];
local_B[col+64] = B[(col_offset+col)*ldb+idx+(2*val_per_iter)];
local_B[col+96] = B[(col_offset+col)*ldb+idx+(3*val_per_iter)];
2023-05-02 19:10:32 +00:00
}
2023-05-02 23:15:38 +00:00
loaded_values = 3;
2023-05-02 19:10:32 +00:00
}
else
{
2023-05-02 23:15:38 +00:00
if(loaded_values == 3)
{
local_A[0] = local_A[1];
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = local_B[col+(32)];
}
else if(loaded_values == 2)
{
local_A[0] = local_A[2];
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = local_B[col+(64)];
}
else
{
local_A[0] = local_A[3];
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = local_B[col+(96)];
}
loaded_values--;
2023-05-02 19:10:32 +00:00
}
2023-05-01 00:43:02 +00:00
2023-05-02 15:58:59 +00:00
smem_A[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*a_tile_offset)] = local_A[0];
2023-05-01 00:43:02 +00:00
2023-05-01 23:38:09 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
2023-05-02 15:58:59 +00:00
smem_B[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*b_tile_offset) + (col*16)] = local_B[col];
2023-05-01 15:21:12 +00:00
}
2023-05-02 14:24:12 +00:00
else if(warp_id < (WARPS-1))
{
local_A[0] = T(0.0);
2023-05-02 15:58:59 +00:00
smem_A[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*a_tile_offset)] = 0.0f;
2023-05-02 14:24:12 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
2023-05-02 15:58:59 +00:00
local_B[col] = 0.0f;
2023-05-02 14:24:12 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
2023-05-02 15:58:59 +00:00
smem_B[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*b_tile_offset) + (col*16)] = 0.0f;
2023-05-02 14:24:12 +00:00
}
2023-05-01 15:21:12 +00:00
ticktock = ticktock == 0 ? 1 : 0;
2023-04-29 00:19:02 +00:00
2023-05-02 14:50:32 +00:00
//for(int base_idx = blockDim.x-32; base_idx < K; base_idx+=blockDim.x-32)
2023-05-02 15:58:59 +00:00
for(int base_idx = blockDim.x-32; base_idx < K; base_idx+=blockDim.x-32)
2023-05-01 15:21:12 +00:00
{
idx = base_idx + threadIdx.x;
2023-05-01 02:12:42 +00:00
2023-05-01 15:21:12 +00:00
__syncthreads();
if(idx < K && warp_id < (WARPS-1))
{
2023-05-02 19:10:32 +00:00
//local_A[0] = A[idx];
2023-05-01 02:12:42 +00:00
2023-05-02 19:10:32 +00:00
//#pragma unroll 32
//for(int col = 0; col < 32; col++)
// local_B[col] = B[(col_offset+col)*ldb+idx];
if(loaded_values == 0)
{
local_A[0] = A[idx];
2023-05-02 23:15:38 +00:00
local_A[1] = A[idx+(1*val_per_iter)];
local_A[2] = A[idx+(2*val_per_iter)];
local_A[3] = A[idx+(3*val_per_iter)];
2023-05-02 19:10:32 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
{
local_B[col] = B[(col_offset+col)*ldb+idx];
2023-05-02 23:15:38 +00:00
local_B[col+32] = B[(col_offset+col)*ldb+idx+(1*val_per_iter)];
local_B[col+64] = B[(col_offset+col)*ldb+idx+(2*val_per_iter)];
local_B[col+96] = B[(col_offset+col)*ldb+idx+(3*val_per_iter)];
2023-05-02 19:10:32 +00:00
}
2023-05-02 23:15:38 +00:00
loaded_values = 3;
2023-05-02 19:10:32 +00:00
}
else
{
2023-05-02 23:15:38 +00:00
if(loaded_values == 3)
{
local_A[0] = local_A[1];
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = local_B[col+(32)];
}
else if(loaded_values == 2)
{
local_A[0] = local_A[2];
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = local_B[col+(64)];
}
else
{
local_A[0] = local_A[3];
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = local_B[col+(96)];
}
loaded_values--;
2023-05-02 19:10:32 +00:00
}
2023-05-01 15:21:12 +00:00
smem_A[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*a_tile_offset)] = local_A[0];
2023-05-01 23:38:09 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
2023-05-01 15:21:12 +00:00
smem_B[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*b_tile_offset) + (col*16)] = local_B[col];
2023-05-01 02:12:42 +00:00
}
2023-05-02 14:24:12 +00:00
else if(warp_id < (WARPS-1))
{
local_A[0] = T(0.0);
smem_A[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*a_tile_offset)] = 0.0f;
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = 0.0f;
#pragma unroll 32
for(int col = 0; col < 32; col++)
smem_B[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*b_tile_offset) + (col*16)] = 0.0f;
}
2023-05-02 15:58:59 +00:00
ticktock = ticktock == 0 ? 1 : 0;
2023-05-01 15:21:12 +00:00
if(warp_id == (WARPS-1))
for(int k = 0; k < batch_size_warps; k++)
{
wmma::load_matrix_sync(a_frag, &(smem_A[(ticktock*batch_size_warps + k)*a_tile_offset]), 16); // 111 mu
wmma::load_matrix_sync(b_frag, &(smem_B[(ticktock*batch_size_warps + k)*b_tile_offset]), 16); // 35 mu
wmma::mma_sync(c_frag, a_frag, b_frag, c_frag);
}
2023-04-29 05:10:42 +00:00
}
2023-04-29 00:19:02 +00:00
2023-05-02 15:58:59 +00:00
__syncthreads();
2023-05-02 18:38:11 +00:00
if(warp_id != (WARPS-1)){ return; }
// only warp_id == (WARPS-1) from here
int warp_lane = threadIdx.x % 32;
2023-05-02 15:58:59 +00:00
ticktock = ticktock == 0 ? 1 : 0;
2023-05-02 18:38:11 +00:00
for(int k = 0; k < batch_size_warps; k++)
{
wmma::load_matrix_sync(a_frag, &(smem_A[(ticktock*batch_size_warps + k)*a_tile_offset]), 16); // 111 mu
wmma::load_matrix_sync(b_frag, &(smem_B[(ticktock*batch_size_warps + k)*b_tile_offset]), 16); // 35 mu
wmma::mma_sync(c_frag, a_frag, b_frag, c_frag);
}
2023-05-02 14:50:32 +00:00
2023-05-01 00:43:02 +00:00
// 129 mu
2023-05-01 15:21:12 +00:00
if(warp_id == (WARPS-1))
2023-05-02 18:38:11 +00:00
wmma::store_matrix_sync(&(smem_A[0]), c_frag, 32, wmma::mem_row_major);
2023-05-02 14:50:32 +00:00
2023-05-02 18:38:11 +00:00
if(col_offset + warp_lane < M)
out[col_offset + warp_lane] = smem_A[warp_lane];
2023-05-31 02:06:59 +00:00
#endif
2023-04-25 23:15:44 +00:00
}
2023-05-31 03:42:21 +00:00
2023-07-04 01:45:38 +00:00
template <typename T> __device__ void printnonzero(T *A, int num_values, const char * strval)
2023-05-31 03:42:21 +00:00
{
for(int i = 0; i < num_values; i++)
if((float)A[i] != 0.0)
2023-07-04 01:45:38 +00:00
printf("%s %i %f\n", strval, i, (float)A[i]);
2023-05-31 03:42:21 +00:00
}
2023-07-04 01:45:38 +00:00
template __device__ void printnonzero<float>(float *A, int num_values, const char*strval);
template __device__ void printnonzero<half>(half *A, int num_values, const char*strval);
2023-05-31 03:42:21 +00:00
2023-05-31 03:07:05 +00:00
__device__ static float nf4_data[16] = {-1.0, -0.6961928009986877, -0.5250730514526367, -0.39491748809814453, -0.28444138169288635, -0.18477343022823334, -0.09105003625154495, 0.0, 0.07958029955625534, 0.16093020141124725, 0.24611230194568634, 0.33791524171829224, 0.44070982933044434, 0.5626170039176941, 0.7229568362236023, 1.0};
2023-04-30 04:52:47 +00:00
template <typename T, int THREADS> __global__ void kgemm_4bit_inference(int M, int N, int K, T * __restrict__ const A, unsigned char *B, float *absmax, T * out, int lda, int ldb, int ldc, int blocksize)
{
2023-07-10 16:53:23 +00:00
#if __CUDA_ARCH__ >= 750
2023-05-31 02:06:59 +00:00
using namespace nvcuda;
2023-05-02 23:15:38 +00:00
int col_offset = blockIdx.x *32;
const int warp_id = threadIdx.x / 32;
2023-07-04 01:45:38 +00:00
const int warp_idx = threadIdx.x % 32;
2023-05-02 23:15:38 +00:00
const int half_warp_id = threadIdx.x / 16;
const int half_warp_lane = threadIdx.x % 16;
const int batch_size_warps = (WARPS-1)*2;
2023-04-30 04:52:47 +00:00
2023-05-31 03:07:05 +00:00
T quant_map[16];
#pragma unroll 16
for(int i = 0; i < 16; i++)
quant_map[i] = nf4_data[i];
2023-07-04 01:45:38 +00:00
//__shared__ T quant_map[16*160];
2023-05-31 03:07:05 +00:00
2023-05-02 23:15:38 +00:00
T local_A[2];
T local_B[64];
unsigned char local_B_4bit[32];
2023-04-30 04:52:47 +00:00
2023-07-04 01:45:38 +00:00
2023-05-02 23:15:38 +00:00
const int a_tile_offset = 16;
const int b_tile_offset = (16*32 + 16);
2023-04-30 04:52:47 +00:00
2023-07-04 01:45:38 +00:00
__shared__ T smem_A[8*16 + (16*(batch_size_warps-1))];
2023-05-02 23:15:38 +00:00
__shared__ T smem_B[2*batch_size_warps*16*32 + (2*16*(batch_size_warps-1))];
2023-07-04 01:45:38 +00:00
__shared__ T smem_C[8*32];
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
wmma::fragment<wmma::matrix_a, 8, 32, 16, half, wmma::row_major> a_frag;
wmma::fragment<wmma::matrix_b, 8, 32, 16, half, wmma::col_major> b_frag;
wmma::fragment<wmma::accumulator, 8, 32, 16, half> c_frag;
wmma::fill_fragment(c_frag, 0.0f);
2023-04-30 04:52:47 +00:00
2023-07-04 01:45:38 +00:00
for(int i = threadIdx.x; i < (8*32); i+=blockDim.x)
smem_C[i] = 0.0f;
__syncthreads();
2023-05-02 23:15:38 +00:00
int ticktock = 0;
int idx = 0 + threadIdx.x;
int loaded_values = 0;
// prefetch
if(idx < K && warp_id < (WARPS-1))
2023-04-30 04:52:47 +00:00
{
2023-05-02 23:15:38 +00:00
if(loaded_values == 0)
{
local_A[0] = A[idx];
local_A[1] = A[idx+blockDim.x-32];
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B_4bit[col] = B[(col_offset+col)*ldb+idx];
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
loaded_values = 1;
}
else
2023-04-30 04:52:47 +00:00
{
2023-05-02 23:15:38 +00:00
local_A[0] = local_A[1];
loaded_values--;
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
#pragma unroll 64
for(int col = 0; col < 64; col+=2)
{
2023-07-04 01:45:38 +00:00
//local_B[col] = dhDequantizeNF4(local_B_4bit[col/2] >> 4)*T(1.0f);
//local_B[col+1] = dhDequantizeNF4(local_B_4bit[col/2] & 0x0F)*T(1.0f);
//local_B[col] = d2DequantizeFP4(local_B_4bit[col/2] >> 4)*(float)(17.0);
//local_B[col+1] = d2DequantizeFP4(local_B_4bit[col/2] & 0x0F)*(float)(17.0);
//local_B[col] = 127*(local_B_4bit[col/2] >> 4)*(float)(17.0);
//local_B[col+1] = 127*(local_B_4bit[col/2] & 0x0F)*(float)(17.0);
//local_B[col] = quant_map[(local_B_4bit[col/2] >> 4)]*T(17.0);
//local_B[col+1] = quant_map[(local_B_4bit[col/2] & 0x0F)]*T(17.0);
local_B[col] = quant_map[160*(local_B_4bit[col/2] >> 4)+warp_idx]*T(17.0);
local_B[col+1] = quant_map[160*(local_B_4bit[col/2] & 0x0F)+warp_idx]*T(17.0);
2023-05-02 23:15:38 +00:00
}
}
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
smem_A[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*a_tile_offset)] = local_A[0];
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
smem_B[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*b_tile_offset) + (col*16)] = local_B[col];
}
else if(warp_id < (WARPS-1))
{
local_A[0] = T(0.0);
smem_A[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*a_tile_offset)] = 0.0f;
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = 0.0f;
#pragma unroll 32
for(int col = 0; col < 32; col++)
smem_B[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*b_tile_offset) + (col*16)] = 0.0f;
}
ticktock = ticktock == 0 ? 1 : 0;
2023-07-04 01:45:38 +00:00
//if(threadIdx.x == 0)
//printf("aa %i %i\n", idx, loaded_values);
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
//for(int base_idx = blockDim.x-32; base_idx < K; base_idx+=blockDim.x-32)
for(int base_idx = blockDim.x-32; base_idx < K; base_idx+=blockDim.x-32)
{
idx = base_idx + threadIdx.x;
2023-07-04 01:45:38 +00:00
//if(threadIdx.x == 0)
//printf("%i %i\n", idx, loaded_values);
2023-05-02 23:15:38 +00:00
2023-07-04 01:45:38 +00:00
//__syncthreads();
2023-05-02 23:15:38 +00:00
if(idx < K && warp_id < (WARPS-1))
{
if(loaded_values == 0)
2023-04-30 04:52:47 +00:00
{
2023-05-02 23:15:38 +00:00
local_A[0] = A[idx];
local_A[1] = A[idx+blockDim.x-32];
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
#pragma unroll 32
for(int col = 0; col < 32; col++)
{
local_B_4bit[col] = B[(col_offset+col)*ldb+idx];
local_B_4bit[col+16] = B[(col_offset+col)*ldb+idx];
}
loaded_values = 1;
}
else
{
local_A[0] = local_A[1];
loaded_values--;
int absidx = (idx + col_offset)/blocksize;
half local_absmax = __ldg(&(absmax[absidx]));
#pragma unroll 64
for(int col = 0; col < 64; col+=2)
{
2023-07-04 01:45:38 +00:00
//local_B[col] = dhDequantizeNF4(local_B_4bit[col/2] >> 4)*T(absidx);
//local_B[col+1] = dhDequantizeNF4(local_B_4bit[col/2] & 0x0F)*T(absidx);
//local_B[col] = T(127)*T(local_B_4bit[col/2] >> 4)*T(absidx);
//local_B[col+1] = T(127)*T(local_B_4bit[col/2] & 0x0F)*T(absidx);
//local_B[col] = quant_map[160*(local_B_4bit[col/2] >> 4)+warp_idx]*T(local_absmax);
//local_B[col+1] = quant_map[160*(local_B_4bit[col/2] & 0x0F)+warp_idx]*T(local_absmax);
local_B[col] = quant_map[(local_B_4bit[col/2] >> 4)]*T(absidx);
local_B[col+1] = quant_map[(local_B_4bit[col/2] & 0x0F)]*T(absidx);
2023-05-02 23:15:38 +00:00
}
2023-07-04 01:45:38 +00:00
//printnonzero<T>(local_B, 128, "");
2023-04-30 04:52:47 +00:00
}
2023-05-02 23:15:38 +00:00
smem_A[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*a_tile_offset)] = local_A[0];
2023-04-30 04:52:47 +00:00
#pragma unroll 32
2023-05-02 23:15:38 +00:00
for(int col = 0; col < 32; col++)
smem_B[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*b_tile_offset) + (col*16)] = local_B[col];
}
else if(warp_id < (WARPS-1))
{
local_A[0] = T(0.0);
smem_A[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*a_tile_offset)] = 0.0f;
#pragma unroll 32
for(int col = 0; col < 32; col++)
local_B[col] = 0.0f;
#pragma unroll 32
for(int col = 0; col < 32; col++)
smem_B[half_warp_lane + (((batch_size_warps*ticktock)+half_warp_id)*b_tile_offset) + (col*16)] = 0.0f;
}
ticktock = ticktock == 0 ? 1 : 0;
if(warp_id == (WARPS-1))
for(int k = 0; k < batch_size_warps; k++)
2023-04-30 04:52:47 +00:00
{
2023-05-02 23:15:38 +00:00
wmma::load_matrix_sync(a_frag, &(smem_A[(ticktock*batch_size_warps + k)*a_tile_offset]), 16); // 111 mu
wmma::load_matrix_sync(b_frag, &(smem_B[(ticktock*batch_size_warps + k)*b_tile_offset]), 16); // 35 mu
wmma::mma_sync(c_frag, a_frag, b_frag, c_frag);
2023-04-30 04:52:47 +00:00
}
}
2023-05-02 23:15:38 +00:00
__syncthreads();
2023-07-04 01:45:38 +00:00
//if(threadIdx.x == 0)
//{
// printnonzero<T>(smem_A, 8*16 + (2*16*(batch_size_warps-1)), "A: ");
// printnonzero<T>(smem_B, 2*batch_size_warps*16*32 + (2*16*(batch_size_warps-1)), "B: ");
//}
2023-05-02 23:15:38 +00:00
if(warp_id != (WARPS-1)){ return; }
// only warp_id == (WARPS-1) from here
int warp_lane = threadIdx.x % 32;
2023-04-30 04:52:47 +00:00
2023-05-02 23:15:38 +00:00
ticktock = ticktock == 0 ? 1 : 0;
for(int k = 0; k < batch_size_warps; k++)
2023-04-30 04:52:47 +00:00
{
2023-07-04 01:45:38 +00:00
//if(warp_lane == 0)
//printf("%i %i %i %i\n", (ticktock*batch_size_warps + k)*a_tile_offset, k, ticktock, threadIdx.x);
2023-05-02 23:15:38 +00:00
wmma::load_matrix_sync(a_frag, &(smem_A[(ticktock*batch_size_warps + k)*a_tile_offset]), 16); // 111 mu
wmma::load_matrix_sync(b_frag, &(smem_B[(ticktock*batch_size_warps + k)*b_tile_offset]), 16); // 35 mu
wmma::mma_sync(c_frag, a_frag, b_frag, c_frag);
2023-04-30 04:52:47 +00:00
}
2023-05-02 23:15:38 +00:00
// 129 mu
if(warp_id == (WARPS-1))
2023-07-04 01:45:38 +00:00
wmma::store_matrix_sync(&(smem_C[0]), c_frag, 32, wmma::mem_row_major);
2023-04-30 04:52:47 +00:00
2023-07-04 01:45:38 +00:00
//printnonzero<T>(smem_C, 32, "");
2023-05-31 03:42:21 +00:00
2023-05-02 23:15:38 +00:00
if(col_offset + warp_lane < M)
2023-07-04 01:45:38 +00:00
out[col_offset + warp_lane] = smem_C[warp_lane];
2023-07-10 16:53:23 +00:00
#endif
2023-04-30 04:52:47 +00:00
}
2023-07-04 22:20:10 +00:00
#define num_values_4bit 32
2023-07-10 04:06:01 +00:00
template <typename T, int THREADS, int BITS> __global__ void kgemm_4bit_inference_naive(int M, int N, int K, T * __restrict__ const A, unsigned char *B, float *absmax, const float *datatype, T * out, int lda, int ldb, int ldc, int blocksize)
2023-07-04 01:45:38 +00:00
{
// per threadblock:
2023-07-11 12:55:49 +00:00
// load step-by-step in chunks of [32,warps]: 1x32 * [32,warps] -> [1,warps]
2023-07-04 01:45:38 +00:00
// 4 warps -> 4 loads per iter
2023-07-11 12:55:49 +00:00
// 1x32 * 32x4 -> 1x4 outputs per thread block
2023-07-08 23:31:58 +00:00
typedef cub::WarpReduce<float> WarpReduce;
2023-07-04 22:20:10 +00:00
__shared__ typename WarpReduce::TempStorage temp_storage[THREADS/32];
2023-07-04 01:45:38 +00:00
const int warp_idx = threadIdx.x / 32;
const int warp_lane = threadIdx.x % 32;
2023-07-04 22:20:10 +00:00
const int row_B = (THREADS/32)*blockIdx.x + warp_idx;
const int num_values_8bit = num_values_4bit/2;
2023-07-08 23:31:58 +00:00
float local_C = 0.0f;
2023-07-08 21:27:12 +00:00
2023-07-04 22:20:10 +00:00
unsigned char local_B_4bit[num_values_8bit];
2023-07-04 01:45:38 +00:00
T local_B[num_values_4bit];
2023-07-04 22:20:10 +00:00
T local_A[num_values_4bit];
2023-07-08 23:31:58 +00:00
__shared__ T quant_map[16];
T local_absmax = T(0.0f);
2023-07-04 01:45:38 +00:00
2023-07-08 23:31:58 +00:00
for(int i = threadIdx.x; i < 16; i++)
2023-07-10 16:53:23 +00:00
quant_map[i] = T(datatype[i]);
2023-07-08 23:31:58 +00:00
__syncthreads();
2023-07-04 01:45:38 +00:00
// A: [1, K]
// B: [N, K]
for(int inner_idx = warp_lane*num_values_4bit; inner_idx < K; inner_idx += 32*num_values_4bit)
{
2023-07-04 22:20:10 +00:00
int inner_idx_halved = inner_idx/2;
int offset_B = ldb*row_B;
int absidx = ((2*offset_B)+inner_idx)/blocksize;
2023-07-08 23:31:58 +00:00
local_absmax = __ldg(&(absmax[absidx]));
2023-07-04 01:45:38 +00:00
2023-07-04 22:20:10 +00:00
if(row_B < M)
2023-07-04 01:45:38 +00:00
{
2023-07-10 20:57:44 +00:00
if((inner_idx_halved + num_values_8bit) < (K/2))
2023-07-04 22:20:10 +00:00
{
2023-07-11 12:55:49 +00:00
// this is the most important for performance considerations
2023-07-04 22:20:10 +00:00
reinterpret_cast<int4(&)[num_values_8bit]>(local_B_4bit)[0] = reinterpret_cast<int4*>(B)[(offset_B+(inner_idx_halved))/(num_values_8bit)];
}
2023-07-04 01:45:38 +00:00
else
2023-07-04 22:20:10 +00:00
{
#pragma unroll
for(int j = 0; j < (num_values_8bit); j++)
2023-07-10 20:57:44 +00:00
if((inner_idx_halved) + j < (K/2))
2023-07-09 19:04:09 +00:00
local_B_4bit[j] = B[offset_B+inner_idx_halved + j];
else
2023-07-04 22:20:10 +00:00
local_B_4bit[j] = 0b01110111;
}
2023-07-04 01:45:38 +00:00
}
2023-07-10 20:57:44 +00:00
else
{
#pragma unroll
for(int j = 0; j < (num_values_8bit); j++)
local_B_4bit[j] = 0b01110111;
}
2023-07-04 01:45:38 +00:00
2023-07-08 23:31:58 +00:00
#pragma unroll
2023-07-10 20:57:44 +00:00
for(int k = 0; k < num_values_8bit; k++)
2023-07-04 01:45:38 +00:00
{
2023-07-10 16:53:23 +00:00
#if __CUDA_ARCH__ >= 800
local_B[k*2] = quant_map[local_B_4bit[k] >> 4]*local_absmax;
local_B[k*2 + 1] = quant_map[local_B_4bit[k] & 0x0F]*local_absmax;
#else
// bf16 multipliation not supported
local_B[k*2] = T((float)quant_map[local_B_4bit[k] >> 4]*(float)local_absmax);
local_B[k*2 + 1] = T((float)quant_map[local_B_4bit[k] & 0x0F]*(float)local_absmax);
#endif
2023-07-04 01:45:38 +00:00
}
2023-07-10 21:24:33 +00:00
if(inner_idx+num_values_4bit < K)
2023-07-04 22:20:10 +00:00
{
2023-07-11 12:55:49 +00:00
// this is also relatively important for performance
2023-07-10 04:06:01 +00:00
if(BITS==16)
{
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[0] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/4) + 0];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[1] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/4) + 1];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[2] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/4) + 2];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[3] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/4) + 3];
}
else
{
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[0] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/8) + 0];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[1] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/8) + 1];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[2] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/8) + 2];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[3] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/8) + 3];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[4] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/8) + 4];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[5] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/8) + 5];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[6] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/8) + 6];
reinterpret_cast<int4(&)[num_values_4bit]>(local_A)[7] = reinterpret_cast<int4*>(A)[inner_idx/(num_values_4bit/8) + 7];
}
2023-07-08 23:31:58 +00:00
2023-07-04 22:20:10 +00:00
}
else
2023-07-11 12:55:49 +00:00
#pragma unroll
2023-07-04 22:20:10 +00:00
for(int k = 0; k < num_values_4bit; k++)
2023-07-10 21:24:33 +00:00
if(inner_idx + k < K)
local_A[k] = A[inner_idx + k];
else
local_A[k] = T(0.0f);
2023-07-04 01:45:38 +00:00
2023-07-11 12:55:49 +00:00
// accumulate in float; small performance hit for Ampere, but lower error for outputs
2023-07-04 01:45:38 +00:00
#pragma unroll
for(int k = 0; k < num_values_4bit; k++)
2023-07-10 16:53:23 +00:00
{
#if __CUDA_ARCH__ >= 800
local_C += (float)(local_A[k]*local_B[k]);
#else
// bf16 multipliation not supported
local_C += ((float)local_A[k]*(float)local_B[k]);
#endif
}
2023-07-04 01:45:38 +00:00
}
local_C = WarpReduce(temp_storage[warp_idx]).Sum(local_C);
if(row_B < M && warp_lane == 0)
2023-07-08 23:31:58 +00:00
out[row_B] = T(local_C);
2023-07-04 01:45:38 +00:00
}
2023-04-29 04:29:40 +00:00
//#define ROWS 2
//template <typename T, int ITEMS, int THREADS> __global__ void gemm_device(int M, int N, int K, T const* A, T* B, T * out, int lda, int ldb, int ldc)
//{
//// 0. We want to fill a 8x128 tile for a thread block so we have 8x16 tile for each warp
//// 1. Load dataB into register
//// 2. Dequantize B
//// 3. Fetch data from A and multiply
//
// typedef cub::BlockLoad<T, THREADS , ITEMS, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadA;
// //__shared__ typename LoadA::TempStorage loada;
// typedef cub::BlockLoad<T, THREADS , ITEMS, cub::BLOCK_LOAD_WARP_TRANSPOSE> LoadB;
// //__shared__ typename LoadB::TempStorage loadb;
// typedef cub::BlockReduce<T, THREADS> BlockReduce;
// // Allocate shared memory for BlockReduce
// //__shared__ typename BlockReduce::TempStorage reduce;
//
// __shared__ union {
// typename BlockReduce::TempStorage reduce;
// typename LoadB::TempStorage loadb;
// typename LoadA::TempStorage loada;
// } temp_storage;
//
//
// T dataA[ITEMS];
// T local_B[ITEMS];
// T local_accC[ROWS];
// int valid_items = 0;
// const int col_offset = blockIdx.x * 8;
//
// __shared__ T tileA[ROWS*THREADS*ITEMS];
// __shared__ T accumulatorC[ROWS*8];
//
// //#pragma unroll 8
// //for(int i = 0; i < 8; i++)
// // tileA[threadIdx.x + (i*256)] = 0.0f;
// //__syncthreads();
// if(threadIdx.x < 64)
// accumulatorC[threadIdx.x] = 0.0f;
// __syncthreads();
//
//
// for(int inner_idx = 0; inner_idx < K; inner_idx+= THREADS*ITEMS)
// {
// valid_items = K - inner_idx > THREADS*ITEMS ? THREADS*ITEMS : K - inner_idx;
// int baserow = 0;
// for(int row = baserow; row < (baserow+ROWS) && row < N; row++)
// {
// LoadA(temp_storage.loada).Load(&(A[(row*K) + inner_idx]), dataA, valid_items, 0.0f);
//
// #pragma unroll ITEMS
// for(int k = 0; k < ITEMS; k++)
// tileA[row*THREADS*ITEMS + threadIdx.x + (k*THREADS)] = dataA[k];
//
// __syncthreads();
// }
// baserow += ROWS;
//
// // load 16 columns from B at a time. B is transposed, so its like loading rows
// // each warp loads one row
// // each thread loads 128 byte
//
// // col: inner_idx + warp_lane
// // row: ldb*(offset + warp_id)
// for(int col = 0; col < 8 && (col_offset + col) < M; col++)
// {
// int colB = col_offset + col;
//
// for(int k = 0; k < ROWS; k++)
// local_accC[k] = 0.0f;
//
// int base_idxB = ldb*colB;
// valid_items = K - inner_idx > THREADS*ITEMS ? THREADS*ITEMS : K - inner_idx;
// LoadB(temp_storage.loadb).Load(&(B[base_idxB + inner_idx]), local_B, valid_items, 0.0f);
// __syncthreads();
//
// for(int row = 0; row < ROWS && row < N; row++)
// {
// #pragma unroll ITEMS
// for(int k = 0; k < ITEMS; k++)
// {
// int idxA = row*THREADS*ITEMS + threadIdx.x + (THREADS*k);
// local_accC[row] += tileA[idxA]*local_B[k];
// }
//
// local_accC[row] = BlockReduce(temp_storage.reduce).Reduce(local_accC[row], cub::Sum());
// if(threadIdx.x == 0)
// atomicAdd(&accumulatorC[row*8 + col], local_accC[row]);
// }
// }
// }
//
// for(int row = 0; row < ROWS && row < N; row++)
// {
// int out_idx = ldc*row + col_offset;
//
// //if(threadIdx.x < 8)
// // if(accumulatorC[row*8 + threadIdx.x] != 0.0)
// // printf("%i %i %i %i %f idx %i %i %i\n", row, col_offset, threadIdx.x, N, accumulatorC[row*8 + threadIdx.x], ldc, out_idx, blockIdx.x);
//
// if(threadIdx.x < 8 && (col_offset + threadIdx.x) < M)
// {
// //printf("%i %i %i %i %f idx %i %i\n", row, col_offset, threadIdx.x, N, accumulatorC[row*8 + threadIdx.x], ldc, out_idx);
// out[out_idx + threadIdx.x] = accumulatorC[row*8 + threadIdx.x];
// }
// }
//
//
//
//}
2023-04-25 23:13:43 +00:00
2023-05-06 18:14:06 +00:00
template <typename T, int FUNC> __global__ void kfunc(T *A, T *B, T value, long n)
2023-04-27 22:12:49 +00:00
{
2023-05-06 18:14:06 +00:00
for(long i = (blockDim.x*blockIdx.x) + threadIdx.x; i < n; i+=(blockDim.x*gridDim.x))
{
switch(FUNC)
{
case FILL:
A[i] = (T)value;
break;
case ARANGE:
A[i] = (T)i;
break;
case _MUL:
A[i] = A[i]*B[i];
break;
2023-04-27 22:12:49 +00:00
}
2023-05-06 18:14:06 +00:00
}
2023-04-27 22:12:49 +00:00
}
2023-04-27 00:12:34 +00:00
2021-10-06 02:16:20 +00:00
//==============================================================
// TEMPLATE DEFINITIONS
//==============================================================
2023-05-06 18:14:06 +00:00
template __global__ void kfunc<float, FILL>(float *A, float *B, float value, long n);
template __global__ void kfunc<unsigned char, FILL>(unsigned char *A, unsigned char *B, unsigned char value, long n);
template __global__ void kfunc<float, ARANGE>(float *A, float *B, float value, long n);
template __global__ void kfunc<float, _MUL>(float *A, float *B, float value, long n);
2023-04-29 05:10:42 +00:00
// these are not used and make no sense, but the compiler needs them
2023-05-01 00:43:02 +00:00
//template __global__ void gemm_device<float, 16, 128>(int M, int N, int K, float * __restrict__ const A, float* B, float * out, int lda, int ldb, int ldc);
2023-05-01 15:21:12 +00:00
template __global__ void gemm_device<half, 32, 256>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-01 23:38:09 +00:00
template __global__ void gemm_device<half, 32, 192>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-02 18:38:11 +00:00
template __global__ void gemm_device<half, 32, 160>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-04-29 05:10:42 +00:00
template __global__ void gemm_device<half, 32, 128>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-01 00:43:02 +00:00
//template __global__ void gemm_device<float, 16, 32>(int M, int N, int K, float * __restrict__ const A, float* B, float * out, int lda, int ldb, int ldc);
template __global__ void gemm_device<half, 32, 32>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-01 01:06:01 +00:00
template __global__ void gemm_device<half, 32, 64>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-01 23:38:09 +00:00
template __global__ void gemm_device<half, 32, 96>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-04-29 05:10:42 +00:00
// these are not used and make no sense, but the compiler needs them
2023-05-01 00:43:02 +00:00
//template __global__ void gemm_device<float, 32, 128>(int M, int N, int K, float * __restrict__ const A, float* B, float * out, int lda, int ldb, int ldc);
2023-05-01 15:21:12 +00:00
template __global__ void gemm_device<half, 16, 256>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-01 23:38:09 +00:00
template __global__ void gemm_device<half, 16, 192>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-02 18:38:11 +00:00
template __global__ void gemm_device<half, 16, 160>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-04-29 04:29:40 +00:00
template __global__ void gemm_device<half, 16, 128>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-01 00:43:02 +00:00
//template __global__ void gemm_device<float, 32, 32>(int M, int N, int K, float * __restrict__ const A, float* B, float * out, int lda, int ldb, int ldc);
template __global__ void gemm_device<half, 16, 32>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-01 01:06:01 +00:00
template __global__ void gemm_device<half, 16, 64>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-05-01 23:38:09 +00:00
template __global__ void gemm_device<half, 16, 96>(int M, int N, int K, half * __restrict__ const A, half* B, half * out, int lda, int ldb, int ldc);
2023-04-27 00:12:34 +00:00
2023-07-04 01:45:38 +00:00
template __global__ void kgemm_4bit_inference<half, 96>(int M, int N, int K, half * __restrict__ const A, unsigned char *B, float *absmax, half * out, int lda, int ldb, int ldc, int blocksize);
2023-04-30 04:52:47 +00:00
template __global__ void kgemm_4bit_inference<half, 128>(int M, int N, int K, half * __restrict__ const A, unsigned char *B, float *absmax, half * out, int lda, int ldb, int ldc, int blocksize);
2023-05-02 23:15:38 +00:00
template __global__ void kgemm_4bit_inference<half, 160>(int M, int N, int K, half * __restrict__ const A, unsigned char *B, float *absmax, half * out, int lda, int ldb, int ldc, int blocksize);
2023-07-04 01:45:38 +00:00
template __global__ void kgemm_4bit_inference<half, 256>(int M, int N, int K, half * __restrict__ const A, unsigned char *B, float *absmax, half * out, int lda, int ldb, int ldc, int blocksize);
2023-07-10 04:06:01 +00:00
template __global__ void kgemm_4bit_inference_naive<half, 128, 16>(int M, int N, int K, half * __restrict__ const A, unsigned char *B, float *absmax, const float *datatype, half * out, int lda, int ldb, int ldc, int blocksize);
template __global__ void kgemm_4bit_inference_naive<__nv_bfloat16, 128, 16>(int M, int N, int K, __nv_bfloat16 * __restrict__ const A, unsigned char *B, float *absmax, const float *datatype, __nv_bfloat16 * out, int lda, int ldb, int ldc, int blocksize);
template __global__ void kgemm_4bit_inference_naive<float, 128, 32>(int M, int N, int K, float * __restrict__ const A, unsigned char *B, float *absmax, const float *datatype, float * out, int lda, int ldb, int ldc, int blocksize);
2023-04-30 04:52:47 +00:00
2022-07-27 00:39:30 +00:00
template __global__ void kExtractOutliers<COL_TURING>(char *A, int *idx, char *out, int idx_size, int rowsA, int colsA, int tiledRowsA, int tiledColsA);
template __global__ void kExtractOutliers<COL_AMPERE>(char *A, int *idx, char *out, int idx_size, int rowsA, int colsA, int tiledRowsA, int tiledColsA);
2022-07-26 19:12:38 +00:00
2022-07-22 21:41:05 +00:00
template __global__ void kspmm_coo_very_sparse_naive<half, 8, 16>(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, half *B, half *out, float *dequant_stats, int nnz, int rowsA, int rowsB, int colsB);
template __global__ void kspmm_coo_very_sparse_naive<half, 16, 16>(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, half *B, half *out, float *dequant_stats, int nnz, int rowsA, int rowsB, int colsB);
template __global__ void kspmm_coo_very_sparse_naive<half, 32, 16>(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, half *B, half *out, float *dequant_stats, int nnz, int rowsA, int rowsB, int colsB);
template __global__ void kspmm_coo_very_sparse_naive<signed char, 8, 8>(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, signed char *B, half *out, float *dequant_stats, int nnz, int rowsA, int rowsB, int colsB);
template __global__ void kspmm_coo_very_sparse_naive<signed char, 16, 8>(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, signed char *B, half *out, float *dequant_stats, int nnz, int rowsA, int rowsB, int colsB);
template __global__ void kspmm_coo_very_sparse_naive<signed char, 32, 8>(int *max_count, int *max_idx, int *offset_rowidx, int *rowidx, int *colidx, half *values, signed char *B, half *out, float *dequant_stats, int nnz, int rowsA, int rowsB, int colsB);
template __global__ void kTransformRowToFormat<256, 8, 32, 32*8, 0, COL32>(char *__restrict__ const A, char *out, int rows, int cols, int tiledCols, int outRows, int outCols);
template __global__ void kTransformRowToFormat<256, 8, 32, 32*8, 1, COL32>(char *__restrict__ const A, char *out, int rows, int cols, int tiledCols, int outRows, int outCols);
template __global__ void kTransformRowToFormat<256, 8, 32, 32*8, 0, COL_TURING>(char *__restrict__ const A, char *out, int rows, int cols, int tiledCols, int outRows, int outCols);
template __global__ void kTransformRowToFormat<256, 8, 32, 32*8, 1, COL_TURING>(char *__restrict__ const A, char *out, int rows, int cols, int tiledCols, int outRows, int outCols);
template __global__ void kTransformRowToFormat<256, 8, 32, 32*8, 0, COL_AMPERE>(char *__restrict__ const A, char *out, int rows, int cols, int tiledCols, int outRows, int outCols);
template __global__ void kTransformRowToFormat<256, 8, 32, 32*8, 1, COL_AMPERE>(char *__restrict__ const A, char *out, int rows, int cols, int tiledCols, int outRows, int outCols);
2022-08-16 17:56:17 +00:00
template __global__ void kdequant_mm_int32_fp16<4, 128, 512>(int *__restrict__ const A, float *__restrict__ const rowStats, float *__restrict__ const colStats, half *out, float* newRowStats, float* newcolStats, half * __restrict__ const bias, const int numRows, const int numCols, const int tileCols, const int n);
2022-07-22 21:41:05 +00:00
template __global__ void kDoubleRowColQuant<64, 4, 16, 64*4, 0>(half *__restrict__ const A, float *__restrict__ const rowStats, float * __restrict__ const colStats, char *out_col_normed, char *out_row_normed, int *rowidx, int *colidx, half *val, int * __restrict__ nnz_block_ptr, float threshold, int rows, int cols, int tiledCols);
template __global__ void kDoubleRowColQuant<64, 4, 16, 64*4, 1>(half *__restrict__ const A, float *__restrict__ const rowStats, float * __restrict__ const colStats, char *out_col_normed, char *out_row_normed, int *rowidx, int *colidx, half *val, int * __restrict__ nnz_block_ptr, float threshold, int rows, int cols, int tiledCols);
2021-10-06 02:16:20 +00:00
template __device__ unsigned char dQuantize<0>(float* smem_code, const float rand, float x);
template __device__ unsigned char dQuantize<1>(float* smem_code, const float rand, float x);
template __global__ void kEstimateQuantiles(float *__restrict__ const A, float *code, const float offset, const float max_val, const int n);
template __global__ void kEstimateQuantiles(half *__restrict__ const A, float *code, const float offset, const half max_val, const int n);
#define MAKE_PreconditionOptimizer32bit1State(oname, gtype) \
template __global__ void kPreconditionOptimizer32bit1State<gtype, oname, 4096, 8>(gtype* g, gtype* p, \
float* state1, float *unorm, \
2023-03-10 21:00:59 +00:00
const float beta1, const float beta2, const float eps, const float weight_decay, \
2021-10-06 02:16:20 +00:00
const int step, const float lr, const float gnorm_scale, const int n); \
MAKE_PreconditionOptimizer32bit1State(MOMENTUM, half)
MAKE_PreconditionOptimizer32bit1State(MOMENTUM, float)
MAKE_PreconditionOptimizer32bit1State(RMSPROP, half)
MAKE_PreconditionOptimizer32bit1State(RMSPROP, float)
2023-03-09 18:10:19 +00:00
MAKE_PreconditionOptimizer32bit1State(LION, half)
MAKE_PreconditionOptimizer32bit1State(LION, float)
2023-05-24 02:37:38 +00:00
MAKE_PreconditionOptimizer32bit1State(LION, __nv_bfloat16)
2021-11-10 23:10:02 +00:00
MAKE_PreconditionOptimizer32bit1State(ADAGRAD, half)
MAKE_PreconditionOptimizer32bit1State(ADAGRAD, float)
2021-10-06 02:16:20 +00:00
#define MAKE_Optimizer32bit1State(oname, gtype) \
template __global__ void kOptimizer32bit1State<gtype, oname>(gtype* g, gtype* p, float* state1, float *unorm, const float max_unorm, const float param_norm, \
2023-03-10 21:00:59 +00:00
const float beta1, const float beta2, const float eps, const float weight_decay,const int step, const float lr, const float gnorm_scale, const bool skip_zeros, const int n); \
2021-10-06 02:16:20 +00:00
MAKE_Optimizer32bit1State(MOMENTUM, half)
MAKE_Optimizer32bit1State(MOMENTUM, float)
MAKE_Optimizer32bit1State(RMSPROP, half)
MAKE_Optimizer32bit1State(RMSPROP, float)
2023-03-09 18:10:19 +00:00
MAKE_Optimizer32bit1State(LION, half)
MAKE_Optimizer32bit1State(LION, float)
2023-05-24 02:37:38 +00:00
MAKE_Optimizer32bit1State(LION, __nv_bfloat16)
2021-11-10 23:10:02 +00:00
MAKE_Optimizer32bit1State(ADAGRAD, half)
MAKE_Optimizer32bit1State(ADAGRAD, float)
2021-10-06 02:16:20 +00:00
#define MAKE_PreconditionOptimizer32bit2State(oname, gtype) \
template __global__ void kPreconditionOptimizer32bit2State<gtype, oname, 4096, 8>(gtype* g, gtype* p, \
float* state1, float* state2, float *unorm, \
const float beta1, const float beta2, const float eps, const float weight_decay, \
const int step, const float lr, const float gnorm_scale, const int n); \
MAKE_PreconditionOptimizer32bit2State(ADAM, float)
2023-04-18 01:01:49 +00:00
MAKE_PreconditionOptimizer32bit2State(ADAM, half)
MAKE_PreconditionOptimizer32bit2State(ADAM, __nv_bfloat16)
2021-10-06 02:16:20 +00:00
2023-04-18 01:01:49 +00:00
template __global__ void kOptimizer32bit2State<float, ADAM>(float* g, float* p, float* state1, float* state2, float *unorm, const float max_unorm, const float param_norm,
const float beta1, const float beta2, const float eps, const float weight_decay,const int step, const float lr, const float gnorm_scale, const bool skip_zeros, const int n);
2021-10-06 02:16:20 +00:00
template __global__ void kOptimizer32bit2State<half, ADAM>(half* g, half* p, float* state1, float* state2, float *unorm, const float max_unorm, const float param_norm,
2021-10-21 01:37:44 +00:00
const float beta1, const float beta2, const float eps, const float weight_decay,const int step, const float lr, const float gnorm_scale, const bool skip_zeros, const int n);
2023-04-18 01:01:49 +00:00
template __global__ void kOptimizer32bit2State<__nv_bfloat16, ADAM>(__nv_bfloat16* g, __nv_bfloat16* p, float* state1, float* state2, float *unorm, const float max_unorm, const float param_norm,
2021-10-21 01:37:44 +00:00
const float beta1, const float beta2, const float eps, const float weight_decay,const int step, const float lr, const float gnorm_scale, const bool skip_zeros, const int n);
2021-10-06 02:16:20 +00:00
#define MAKE_PreconditionStatic8bit1State(oname, gtype) \
template __global__ void kPreconditionOptimizerStatic8bit1State<gtype, oname>(gtype* p, gtype* __restrict__ const g, unsigned char*__restrict__ const state1, \
float *unorm, \
const float beta1, \
2023-03-10 21:00:59 +00:00
const float beta2, \
2021-10-06 02:16:20 +00:00
const float eps, const int step, \
float* __restrict__ const quantiles1, \
float* max1, float* new_max1, \
const float weight_decay, \
const float gnorm_scale, \
const int n); \
MAKE_PreconditionStatic8bit1State(MOMENTUM, half)
MAKE_PreconditionStatic8bit1State(MOMENTUM, float)
MAKE_PreconditionStatic8bit1State(RMSPROP, half)
MAKE_PreconditionStatic8bit1State(RMSPROP, float)
2023-03-09 18:10:19 +00:00
MAKE_PreconditionStatic8bit1State(LION, half)
MAKE_PreconditionStatic8bit1State(LION, float)
2021-10-06 02:16:20 +00:00
#define MAKE_optimizerStatic8bit1State(oname, gtype) \
template __global__ void kOptimizerStatic8bit1State<gtype, oname>(gtype* p, gtype* const g, unsigned char* state1, \
const float *unorm, const float max_unorm, const float param_norm, \
const float beta1, \
2023-03-10 21:00:59 +00:00
const float beta2, \
2021-10-06 02:16:20 +00:00
const float eps, const int step, const float lr, \
float* __restrict__ const quantiles1, \
float* max1, float* new_max1, \
float weight_decay, \
const float gnorm_scale, \
const int n); \
MAKE_optimizerStatic8bit1State(MOMENTUM, half)
MAKE_optimizerStatic8bit1State(MOMENTUM, float)
MAKE_optimizerStatic8bit1State(RMSPROP, half)
MAKE_optimizerStatic8bit1State(RMSPROP, float)
2023-03-09 18:10:19 +00:00
MAKE_optimizerStatic8bit1State(LION, half)
MAKE_optimizerStatic8bit1State(LION, float)
2021-10-06 02:16:20 +00:00
#define MAKE_PreconditionStatic8bit2State(oname, gtype) \
template __global__ void kPreconditionOptimizerStatic8bit2State<gtype, oname>(gtype* p, gtype* __restrict__ const g, unsigned char*__restrict__ const state1, unsigned char* __restrict__ const state2, \
float *unorm, \
const float beta1, const float beta2, \
const float eps, const int step, \
float* __restrict__ const quantiles1, float* __restrict__ const quantiles2, \
float* max1, float* max2, float* new_max1, float* new_max2, \
const float gnorm_scale, \
const int n); \
MAKE_PreconditionStatic8bit2State(ADAM, half)
MAKE_PreconditionStatic8bit2State(ADAM, float)
#define MAKE_optimizerStatic8bit2State(oname, gtype) \
template __global__ void kOptimizerStatic8bit2State<gtype, oname>(gtype* p, gtype* const g, unsigned char* state1, unsigned char* state2, \
const float *unorm, const float max_unorm, const float param_norm, \
const float beta1, const float beta2, \
const float eps, const int step, const float lr, \
float* __restrict__ const quantiles1, float* __restrict__ const quantiles2, \
float* max1, float* max2, float* new_max1, float* new_max2, \
float weight_decay, \
const float gnorm_scale, \
const int n); \
MAKE_optimizerStatic8bit2State(ADAM, half)
MAKE_optimizerStatic8bit2State(ADAM, float)
template __global__ void kPercentileClipping<float, 2048, 4>(float * __restrict__ g, float *gnorm_vec, int step, const int n);
template __global__ void kPercentileClipping<half, 2048, 4>(half * __restrict__ g, float *gnorm_vec, int step, const int n);
2023-04-02 23:10:35 +00:00
#define MAKE_kQuantizeBlockwise(dtype, blocksize, num_per_thread, stochastic, data_type_name) \
template __global__ void kQuantizeBlockwise<dtype, blocksize, num_per_thread, stochastic, data_type_name>(float * code, dtype * __restrict__ const A, float *absmax, unsigned char *out, float * __restrict__ const rand, const int rand_offset, const int n); \
MAKE_kQuantizeBlockwise(half, 4096, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(half, 4096, 4, 1, General8bit)
MAKE_kQuantizeBlockwise(half, 2048, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(half, 1024, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(half, 512, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(half, 256, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(half, 128, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(half, 64, 2, 0, General8bit)
2023-07-05 02:58:31 +00:00
MAKE_kQuantizeBlockwise(half, 4096, 4, 0, FP4)
MAKE_kQuantizeBlockwise(half, 2048, 4, 0, FP4)
MAKE_kQuantizeBlockwise(half, 1024, 4, 0, FP4)
MAKE_kQuantizeBlockwise(half, 512, 2, 0, FP4)
MAKE_kQuantizeBlockwise(half, 256, 2, 0, FP4)
MAKE_kQuantizeBlockwise(half, 128, 2, 0, FP4)
MAKE_kQuantizeBlockwise(half, 64, 2, 0, FP4)
MAKE_kQuantizeBlockwise(half, 4096, 4, 0, NF4)
MAKE_kQuantizeBlockwise(half, 2048, 4, 0, NF4)
MAKE_kQuantizeBlockwise(half, 1024, 4, 0, NF4)
MAKE_kQuantizeBlockwise(half, 512, 2, 0, NF4)
MAKE_kQuantizeBlockwise(half, 256, 2, 0, NF4)
MAKE_kQuantizeBlockwise(half, 128, 2, 0, NF4)
MAKE_kQuantizeBlockwise(half, 64, 2, 0, NF4)
2023-04-02 23:10:35 +00:00
MAKE_kQuantizeBlockwise(float, 4096, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(float, 4096, 4, 1, General8bit)
MAKE_kQuantizeBlockwise(float, 2048, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(float, 1024, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(float, 512, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(float, 256, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(float, 128, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(float, 64, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(float, 4096, 4, 0, FP4)
MAKE_kQuantizeBlockwise(float, 2048, 4, 0, FP4)
MAKE_kQuantizeBlockwise(float, 1024, 4, 0, FP4)
MAKE_kQuantizeBlockwise(float, 512, 2, 0, FP4)
MAKE_kQuantizeBlockwise(float, 256, 2, 0, FP4)
MAKE_kQuantizeBlockwise(float, 128, 2, 0, FP4)
MAKE_kQuantizeBlockwise(float, 64, 2, 0, FP4)
MAKE_kQuantizeBlockwise(float, 4096, 4, 0, NF4)
MAKE_kQuantizeBlockwise(float, 2048, 4, 0, NF4)
MAKE_kQuantizeBlockwise(float, 1024, 4, 0, NF4)
MAKE_kQuantizeBlockwise(float, 512, 2, 0, NF4)
MAKE_kQuantizeBlockwise(float, 256, 2, 0, NF4)
MAKE_kQuantizeBlockwise(float, 128, 2, 0, NF4)
MAKE_kQuantizeBlockwise(float, 64, 2, 0, NF4)
2023-07-05 02:58:31 +00:00
MAKE_kQuantizeBlockwise(__nv_bfloat16, 4096, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 4096, 4, 1, General8bit)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 2048, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 1024, 4, 0, General8bit)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 512, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 256, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 128, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 64, 2, 0, General8bit)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 4096, 4, 0, FP4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 2048, 4, 0, FP4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 1024, 4, 0, FP4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 512, 2, 0, FP4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 256, 2, 0, FP4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 128, 2, 0, FP4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 64, 2, 0, FP4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 4096, 4, 0, NF4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 2048, 4, 0, NF4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 1024, 4, 0, NF4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 512, 2, 0, NF4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 256, 2, 0, NF4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 128, 2, 0, NF4)
MAKE_kQuantizeBlockwise(__nv_bfloat16, 64, 2, 0, NF4)
2023-04-02 23:10:35 +00:00
template __global__ void kDequantizeBlockwise<half, 512, 64, 8, FP4>(float *code, unsigned char * A, float * absmax, half *out, const int blocksize, const int n);
template __global__ void kDequantizeBlockwise<half, 512, 64, 8, General8bit>(float *code, unsigned char * A, float * absmax, half *out, const int blocksize, const int n);
template __global__ void kDequantizeBlockwise<half, 512, 64, 8, NF4>(float *code, unsigned char * A, float * absmax, half *out, const int blocksize, const int n);
2023-07-05 02:58:31 +00:00
template __global__ void kDequantizeBlockwise<float, 512, 64, 8, FP4>(float *code, unsigned char * A, float * absmax, float *out, const int blocksize, const int n);
template __global__ void kDequantizeBlockwise<float, 512, 64, 8, General8bit>(float *code, unsigned char * A, float * absmax, float *out, const int blocksize, const int n);
2023-04-02 23:10:35 +00:00
template __global__ void kDequantizeBlockwise<float, 512, 64, 8, NF4>(float *code, unsigned char * A, float * absmax, float *out, const int blocksize, const int n);
2023-07-05 02:58:31 +00:00
template __global__ void kDequantizeBlockwise<__nv_bfloat16, 512, 64, 8, FP4>(float *code, unsigned char * A, float * absmax, __nv_bfloat16 *out, const int blocksize, const int n);
template __global__ void kDequantizeBlockwise<__nv_bfloat16, 512, 64, 8, General8bit>(float *code, unsigned char * A, float * absmax, __nv_bfloat16 *out, const int blocksize, const int n);
template __global__ void kDequantizeBlockwise<__nv_bfloat16, 512, 64, 8, NF4>(float *code, unsigned char * A, float * absmax, __nv_bfloat16 *out, const int blocksize, const int n);
2021-10-06 02:16:20 +00:00
#define MAKE_OptimizerStatic8bit2StateBlockwise(oname, gtype, block_size, num_per_thread) \
template __global__ void kOptimizerStatic8bit2StateBlockwise<gtype, oname, block_size, num_per_thread>(gtype* p, gtype* __restrict__ const g, unsigned char* state1, unsigned char* state2, \
const float beta1, const float beta2, \
const float eps, const int step, const float lr, \
float* __restrict__ const quantiles1, float* __restrict__ const quantiles2, \
float* absmax1, float* absmax2, \
float weight_decay, \
2021-10-21 01:37:44 +00:00
const float gnorm_scale, const bool skip_zeros, const int n); \
2021-10-06 02:16:20 +00:00
MAKE_OptimizerStatic8bit2StateBlockwise(ADAM, float, 2048, 8)
MAKE_OptimizerStatic8bit2StateBlockwise(ADAM, half, 2048, 8)
2023-04-01 17:33:03 +00:00
MAKE_OptimizerStatic8bit2StateBlockwise(ADAM, __nv_bfloat16, 2048, 8)
2021-10-06 02:16:20 +00:00
#define MAKE_OptimizerStatic8bit1StateBlockwise(oname, gtype, block_size, num_per_thread) \
template __global__ void kOptimizerStatic8bit1StateBlockwise<gtype, oname, block_size, num_per_thread>( \
gtype* p, gtype* __restrict__ const g, unsigned char* state1, \
const float beta1, const float beta2, \
const float eps, const int step, const float lr, \
float* __restrict__ const quantiles1, \
float* absmax1, \
float weight_decay, \
2021-10-21 01:37:44 +00:00
const float gnorm_scale, const bool skip_zeros, const int n); \
2021-10-06 02:16:20 +00:00
MAKE_OptimizerStatic8bit1StateBlockwise(MOMENTUM, float, 2048, 8)
MAKE_OptimizerStatic8bit1StateBlockwise(MOMENTUM, half, 2048, 8)
MAKE_OptimizerStatic8bit1StateBlockwise(RMSPROP, float, 2048, 8)
MAKE_OptimizerStatic8bit1StateBlockwise(RMSPROP, half, 2048, 8)
2023-03-09 18:10:19 +00:00
MAKE_OptimizerStatic8bit1StateBlockwise(LION, float, 2048, 8)
MAKE_OptimizerStatic8bit1StateBlockwise(LION, half, 2048, 8)
2023-05-24 02:37:38 +00:00
MAKE_OptimizerStatic8bit1StateBlockwise(LION, __nv_bfloat16, 2048, 8)
2021-11-10 23:10:02 +00:00
MAKE_OptimizerStatic8bit1StateBlockwise(ADAGRAD, float, 2048, 8)
MAKE_OptimizerStatic8bit1StateBlockwise(ADAGRAD, half, 2048, 8)
